THESIS DESIGN AND CONSTRUCTION OF ELECTRIC MOTOR DYNAMOMETER AND GRID ATTACHED STORAGE LABORATORY Submitted by Markus Lutz Department of Mechanical Engineering In partial fulfillment of the requirements For the Degree of Master of Science Colorado State University Fort Collins, Colorado Fall 2011 Master’s Committee: Advisor: Thomas Bradley Daniel Zimmerle Peter Young Copyright by Markus Lutz 2011 All Rights Reserved ABSTRACT DESIGN AND CONSTRUCTION OF ELECTRIC MOTOR DYNAMOMETER AND GRID ATTACHED STORAGE LABORATORY The purpose of this thesis is to describe the design and development of a laboratory facility to both educate students on electric vehicle components as well as allow researchers to gain experimental results of grid‐attached‐storage testing. With the anticipated roll out of millions of electric vehicles, manufacturers of such vehicles need educated hires with field experience. Through instruction with this lab, Colorado State University plans to be a major resource in equipping the future electric vehicle work force with necessary training and hands‐on experience using real world, full‐scale, automotive grade electric vehicle components. The lab also supports research into grid‐
attached‐storage. This thesis explains the design objectives, challenges, selections, construction and initial testing of the lab, and also provides context for the types of education and research which can be performed utilizing the laboratory. ii ACKNOWLEDGEMENTS This material is based on work supported by the Department of Energy under Award Number DE‐EE0002627. This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe on privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. iii THANKS I would like to take this opportunity to thank everyone who made this research possible. Dr. Thomas Bradley and Daniel Zimmerle provided countless hours of guidance and support. They both have been critical in my graduate research and my development as an engineer. I sincerely thank them for all of the time and support they have given me. This research would not have been possible without the contributions of the American Recovery and Reinvestment Act, Clean Energy Super Cluster, Engineering Student Technology Fee Committee, Colorado State University, and the Engines and Energy Conversion Laboratory (EECL). Phil Bacon, Jacob Renquist, Andrew Geck, Brian Huff and Justin Wagner, supplied a significant amount of technical expertise and assistance in the construction of the lab and for them I am very thankful. I would also like to thank God, my family and my friends who continually provided for me and showed support and encouragement to overcome the many obstacles I faced as well as for helping me pursue my dreams as an engineer. iv TABLE OF CONTENTS ABSTRACT ......................................................................................................................................... ii ACKNOWLEDGEMENTS ................................................................................................................... iii THANKS ........................................................................................................................................... iv TABLE OF CONTENTS........................................................................................................................ v LIST OF FIGURES ............................................................................................................................ viii LIST OF TABLES ............................................................................................................................... xii LIST OF ABBREVIATIONS ................................................................................................................ xiv 1. 2. Introduction to the Electric Drivetrain Teaching Center ......................................................... 1 1.1 Motivation for Education ........................................................................................... 1 1.2 Motivation for Grid‐Attached‐Storage Research ....................................................... 3 1.3 Modes of Education for Dynamometer Laboratory ................................................... 5 1.4 Modes of Research for Grid‐Attached‐Storage .......................................................... 6 1.5 Objectives of the eDTC ............................................................................................... 7 Grid‐Attached‐Storage ........................................................................................................... 10 2.1 Grid‐Attached‐Storage Background ......................................................................... 10 2.1.1 Ancillary Services ...................................................................................................... 10 2.1.2 Energy Arbitrage ...................................................................................................... 12 2.1.3 Renewable Energy Firming ....................................................................................... 14 2.1.4 Second Use Batteries ............................................................................................... 15 2.1.5 Islanded Power Grids, Expeditionary Military, or Emergency Uses ......................... 16 2.2 System Overview ...................................................................................................... 17 2.3 Challenges to Overcome .......................................................................................... 19 2.3.1 Comply with Standards ............................................................................................ 19 2.3.2 Unusual Cooling Requirements for Industrial Labs .................................................. 22 2.3.3 Voltage Mismatch .................................................................................................... 23
v 2.3.4 2.4 Design ....................................................................................................................... 28 2.4.1 Inverter/Variable Frequency Drive .......................................................................... 28 2.4.2 Active Front End and Line Sync Card ........................................................................ 30 2.4.3 Battery Pack ............................................................................................................. 30 2.4.4 Transformer ............................................................................................................. 33 2.4.5 LC Filter ..................................................................................................................... 34 2.4.6 Cooling Circuit .......................................................................................................... 36 2.4.7 Voltage and Frequency Sensor ................................................................................. 37 2.4.8 Unique Safety Requirements ................................................................................... 39 2.4.8.1 Insulation Monitoring Device .......................................................................... 41 2.4.8.2 Insulated Connectors ....................................................................................... 42 2.5 3. Time Delay Impact on Voltage/Frequency Measuring Device ................................. 24 Check Out and Verification ...................................................................................... 43 2.5.1 Finalized System Construction ................................................................................. 43 2.5.2 System Operation ..................................................................................................... 46 2.5.3 Testing Results ......................................................................................................... 48 Electric Vehicle Component Education through a Dynamometer ......................................... 51 3.1 Dynamometer Introduction ..................................................................................... 51 3.1.1 Various Methods of Motor Testing To Date ............................................................ 52 3.1.2 Uniqueness of this Set‐Up ........................................................................................ 53 3.2 High‐level System Overview of the Dynamometer Lab ........................................... 53 3.3 Design Challenges .................................................................................................... 55 3.3.1 Cogging and Other Vibration Effects ........................................................................ 55 3.3.2 Sizing of DC Power Supply ........................................................................................ 56 3.4 Design of this Equipment ......................................................................................... 59 3.4.1 Electric Motors ......................................................................................................... 60 3.4.2 Variable Frequency Drive ......................................................................................... 63 3.4.3 DC Power Supply ...................................................................................................... 63 3.4.4 Current Sensor ......................................................................................................... 65 3.4.5 Torque Sensor and Coupler Selection ...................................................................... 68 3.4.5.1 Objectives ........................................................................................................ 68 3.4.5.2 Method ............................................................................................................ 68 vi 3.4.5.3 Results ............................................................................................................. 72 3.4.5.4 Torque Sensor Selection .................................................................................. 73 3.4.6 Mounting Box for Torque Sensor and Two Electric Motors..................................... 75 3.4.7 Powertrain Control Module ..................................................................................... 81 3.4.8 Control Computer .................................................................................................... 81 3.4.9 Data Acquisition System ........................................................................................... 81 3.4.10 3.5 Check Out and Verification ...................................................................................... 84 3.5.1 Implementation and Construction ........................................................................... 85 3.5.2 Motor Spin‐Up and Coast Down Operation ............................................................. 88 3.6 4. Emergency Stop Circuit ........................................................................................ 83 Course Work Examples ............................................................................................ 89 3.6.1 Lab #1: Map Torque versus Speed Curve ................................................................. 90 3.6.2 Lab #2: Map Motor Efficiency Lab............................................................................ 94 3.6.3 Lab #3: Investigate Temperature Effects on Electric Motors ................................... 99 3.6.4 Lab #4: Intro to Grid‐Attached‐Storage Lab ........................................................... 101 Summary and Conclusions ................................................................................................... 105 4.1 Future of the eDTC ................................................................................................. 105 WORKS CITED ............................................................................................................................... 107 Appendix A – Wiring Schematics ................................................................................................. 113 Appendix B – Artisan Vehicle Systems Pin‐Out Listings ............................................................... 128 Appendix C – GD&T Print for Motor Mounting Box .................................................................... 133 Appendix D –Natural Frequency MatLab Procedure ................................................................... 134 Appendix E – Instructions for Changing the Absorber Inverter from Line Sync Control to Resolver Control (GAS to Dyno Set‐Up) ...................................................................................................... 151 vii LIST OF FIGURES Figure 2‐1: Simplified Layout of Grid‐Attached‐Storage Equipment ................................ 18 Figure 2‐2: Liquid Cooling System Layout ......................................................................... 22 Figure 2‐3: Control Loop for GAS Systems ........................................................................ 25 Figure 2‐4: Impact of Measurement Delay on Microgrid Performance ........................... 26 Figure 2‐5: Detailed Highlights of Battery Pack Components .......................................... 32 Figure 2‐6: Cell Monitors for Each Cell of the Battery Pack.............................................. 33 Figure 2‐7: LC Filter Cabinet with Pre‐Charge Circuit Schematic ...................................... 36 Figure 2‐8: Cooling Pumps with Priming Valves, Flow Meters, and Foot Valves ............. 37 Figure 2‐9: Cooling Circuit Fans, Radiators, and Reservoir ............................................... 37 Figure 2‐10: Caddock Electronics Voltage Divider ............................................................ 39 Figure 2‐11: Bender IR155 Ground Fault Circuit Interrupter ............................................ 41 Figure 2‐12: Cam Connectors by Marinco. Clockwise: Female Cam Connector, Male Cam Connector, and Demonstration of Installation, Panel Mounts ........................................ 42 Figure 2‐13: Overview of Components and Their Locations in 3D CAD Model ................ 44 Figure 2‐14: Actual Layout of GAS Components ............................................................... 45 Figure 2‐15: Actual Layout LC Filter Cabinet ..................................................................... 46 Figure 2‐16: Desk Control Box with Switch Callouts ......................................................... 47 Figure 2‐17: Display Parameters in DSE Lite with Arbitrary Values for Display Purposes Only ................................................................................................................................... 48 Figure 2‐18: Screenshot of DSE Lite in Background with PCM GUI BMS Monitor in Foreground ....................................................................................................................... 49
viii Figure 2‐19: PCM GUI BMS Plot of Pack Current and Voltage vs. Time Plot (left) and Max and Min Cell Voltage vs. Time Plot (right) ........................................................................ 49 Figure 2‐20: Voltage and VAR vs. Time Plot for DC Voltage Command and Real as well as Reactive Power during the Discharge Event ..................................................................... 50 Figure 3‐1: Simplified Diagram of Dyno Set‐Up ................................................................ 54 Figure 3‐2: Schematic for Efficiency Calculations ............................................................. 57 Figure 3‐3: Schematic for eDTC Dynamometer Set‐Up .................................................... 59 Figure 3‐4: Output Torque of GP150WC Motor from Artisan Vehicle Systems ............... 61 Figure 3‐5: Output Power of GP150WC Motor from Artisan Vehicle Systems ................ 62 Figure 3‐6: Efficiency Curve of GP150WC Motor from Artisan Vehicle Systems ............. 62 Figure 3‐7: DC Disconnect Internals Showing Diode Installation ..................................... 64 Figure 3‐8: DC Bus Cabinet with Current Sensors Shown on Right .................................. 67 Figure 3‐9: Model Definitions Used to Determine System Natural Frequency with Mass Numbers Labeled above Different Components .............................................................. 71 Figure 3‐10: Amplitude at each Different Mass Number for the Three Different Modes 72 Figure 3‐11: Visual Representation of Senor and Coupling Cost vs. Natural Frequency Comparison with a Dashed Red Line Representing the Motor Natural Frequency ......... 73 Figure 3‐12: R+W America BK2 Bellows Coupling with Clamping Hub ............................. 74 Figure 3‐13: Sensor Technologies RWT320 Torque Sensor .............................................. 74 Figure 3‐14: Mounting Design Concept A. Motors Mounted on Reinforced Stands with Cantilevered L‐Bracket Supports ...................................................................................... 76 Figure 3‐15: Mounting Design Concept B. Square Box Mounted within Square Plate with Motor Mounting Holes ..................................................................................................... 77 Figure 3‐16: Mounting Design Concept C. Square Box Rotated 45 Degrees to Allow for Most Interior Volume for Sensor and Couplers, Yet Still within Motor Mounting Holes 77 Figure 3‐17: Fundamental Finite Element Analysis of Design C with Concentrated Forces at the Motor Mounting Bolt Holes ................................................................................... 78 ix Figure 3‐18: CAD Model Showing Dimensions between Motor Mounting Bolt Hole and the Next Surface ............................................................................................................... 79 Figure 3‐19: 3D CAD Model of Design C with Mounting Feet Brackets and Bushings ..... 80 Figure 3‐20: Simplistic Schematic of E‐stop Logic ............................................................. 84 Figure 3‐21: 3D CAD Model Detailed Layout with Component Call‐Outs ........................ 85 Figure 3‐22: Inverter Rack with DC Bus Bar Cabinet ......................................................... 86 Figure 3‐24: Dyno Rig with Key Components Called Out ................................................. 87 Figure 3‐25: Control Cabinet Layout ................................................................................. 87 Figure 3‐26: Physical Layout of Constructed Components ............................................... 88 Figure 3‐27: Torque Curve at Continuous and Peak Torque ............................................ 91 Figure 3‐28: PV Simulator Screenshot of Manual_Voltage.vi ........................................... 93 Figure 3‐29: Screenshot and Explanations of Motor_Map.vi ........................................... 93 Figure 3‐30: Flow Schematic with Efficiencies at Each Component ................................. 95 Figure 3‐31: Example Motor Efficiency Map .................................................................... 96 Figure 5‐1: Control Cabinet Wiring ................................................................................. 114 Figure 5‐2: DAQ Wiring ................................................................................................... 115 Figure 5‐3: Linear Power Supply Wiring ......................................................................... 116 Figure 5‐4: Powertrain Control Module Wiring .............................................................. 117 Figure 5‐5: E‐Stop, Breakers, and Light Tower Wiring .................................................... 118 Figure 5‐6: GFCI and Battery Pack Wiring ....................................................................... 119 Figure 5‐7: Control Desk and Computer Wiring ............................................................. 120 Figure 5‐8: GAS Set‐Up Wiring ........................................................................................ 121 Figure 5‐9: Dyno Set‐Up Wiring ...................................................................................... 122 Figure 5‐10: Eagle 500W CPU Power Supply Wiring ....................................................... 123 x Figure 5‐11: Electric Fans Wiring .................................................................................... 124 Figure 5‐12: Terminal Block Layout in Control Cabinet .................................................. 125 Figure 5‐13: Terminal Block Layout in Control Cabinet (continued) .............................. 126 Figure 5‐14: Wiring Bundle Distribution in Control Cabinet ........................................... 127 Figure 5‐15: Natural Frequency Amplitude Response Plot ............................................ 135 Figure 5‐16: Spring‐mass diagram for the Sensor Tech and R+W couplings set‐up ....... 135 Figure 5‐17: Caveat #1 Mode Amplitudes for Each Mass in the Spring‐Mass System ... 141 Figure 5‐18: Caveat #2 Mode Amplitudes of Each Mass ................................................ 141 Figure 5‐19: Mass‐Spring Diagram of Interface Torque Sensor with Caveat #1 Shown (Coupler and Torque Sensor Shaft Springs Paired Together) ......................................... 142 Figure 5‐20: Detailed View of Caveat #1 with Only the First Three Modes Shown ....... 145 Figure 5‐21: Mode Amplitude Plot of Caveat #2 for Interface Sensor ........................... 145 Figure 5‐22: Mode‐Amplitude Plot of Caveat #1 with the Interface Sensor and the Supplied Interface Coupling ............................................................................................ 146 Figure 5‐23: Mode‐Amplitude Plot for Caveat #1 with Interface Sensor and BKM Couplings ......................................................................................................................... 147 Figure 5‐24: Binsfeld Torque Sensors ............................................................................. 148 Figure 5‐25: Spring‐mass diagram of Binsfeld torque sensor set‐up ............................. 148 xi LIST OF TABLES Table 2‐1: Summary of Different Ancillary Services ......................................................... 11 Table 2‐2: Current Distortion Limits for General Distribution Systems (ISC and IL are defined as system short circuit and load size respectively) ............................................. 21 Table 2‐3: Voltage Distortion Limits ................................................................................. 21 Table 2‐4: VFD Selection Comparison Chart ..................................................................... 29 Table 2‐5: GBS 40Ah Battery Cell Specifications ............................................................... 31 Table 2‐6: Jefferson Electric 150kVA Transformer Specifications .................................... 34 Table 2‐7: Voltage/Frequency Sensor Comparison .......................................................... 38 Table 3‐1: Required Make‐Up Power by the Power Supply Calculations ......................... 58 Table 3‐2: Compilation of Torque Sensors Considered for Selection ............................... 69 Table 3‐3: Coupling Selection Values ................................................................................ 70 Table 3‐4: Design Comparison for Different Mounting Box Designs ................................ 76 Table 5‐1: PCM 35 Pin Assignments ............................................................................... 129 Table 5‐2: PCM 25 Pin Assignments ............................................................................... 130 Table 5‐3: Absorber Inverter Feedback Cable Assignments for Dyno and GAS Set‐Ups 130 Table 5‐4: UUT Feedback Cable Pin Assignments ........................................................... 131 Table 5‐5: Deutsche Pin Battery Connector Assignments .............................................. 132 Table 5‐6: Colored Bundle Assignments and CAN Labeling ............................................ 132 Table 5‐7: Sensor Tech with R+W Couplers Caveat Comparison Results ....................... 140 Table 5‐8: Interface Sensor with R+W Couplers Caveat Comparison............................. 144 xii Table 5‐10: Interface coupling comparison results (all using caveat#1) ........................ 146 Table 5‐12: Binsfeld natural frequency results ............................................................... 149 Table 5‐14: Summary of Natural Frequency of Different Torque Sensors and Couplings (lowest reported natural frequency shown for each case) ............................................ 150 xiii LIST OF ABBREVIATIONS AVS ‐ Artisan Vehicle Systems AE ‐ Advanced Energy AFE ‐ Active Front End BMS ‐ Battery Management System CAN Controller Area Network ‐ cRIO ‐ Compact RIO Data Acquisition Chassis CSU Colorado State University ‐ DAQ ‐ Data Acquisition Dyno ‐ Dynamometer eDTC ‐ Electric Drivetrain Teaching Center EMF ‐ Electromotive Force EVs ‐ Electric Vehicles FPGA ‐ Field‐Programmable Gate Array GAS Grid‐Attached‐Storage ‐ GFCI ‐ Ground Fault Circuit Interrupter GUI ‐ Graphical User Interface ICE ‐ Internal Combustion Engine PCC ‐ Point of Common Coupling PCM ‐ Powertrain Control Module POI Point of Interference
‐ xiv SAW ‐ Surface Acoustic Waves SOC ‐ State of Charge UUT ‐ Unit Under Test V2G ‐ Vehicle to Grid VFD ‐ Variable Frequency Drive xv 1.
Introduction to the Electric Drivetrain Teaching Center The purpose of this section of the paper is to describe the motivation behind the Electric Drivetrain Teaching Center (eDTC) for both the educational and the research components of the lab, the research possible in the lab, and finally the objectives for both the drive‐train teaching configuration and grid‐attached‐storage (GAS) configuration of the lab. 1.1 Motivation for Education In the U.S. there are over 137 million passenger vehicles on the road using the finite source of fossil fuel available on Earth. The United States of America has plans to reduce this heavy dependence on fossil fuels, particularly on liquid fossil fuels (Bureau of Transportation: Statistics, 2010)
. President Obama proposes to have 1 million hybrid vehicles on the road by 2015 in order to reduce these emissions and the country’s dependence on foreign oil (Mitlitski, 2011). Additionally, the American Recovery and Reinvestment Act supported increased growth in the electric vehicle sector. Due to these initiatives, the need for educated and trained students by automotive manufactures and clean technology companies is increasing. Ford plans to hire 7,000 new employees to manufacture hybrid and electric vehicles and BMW seeks to hire 2,600 new employees for hybrid and electric system designs (Naughton, 2011) (Reiter, 2010). With these anticipated job openings, the demand for students educated in hybrid vehicle technologies is increasing 1 substantially. However, educational research has shown that a key component to successful engineering students is hands‐on experience (Warrington, Kirkpatrick, & Danielson, 2010). Therefore, there is a need for a facility to provide real world training for students with electric vehicle technology. Research has also shown that “…students have indicated that using testing stations aids significantly in their understanding” (Slocombe, Wagner, Heber, & Harner, 1990)
. Currently, most vehicle drivetrain test systems are designed to test conventional, that is internal combustion engines, drivetrains, and do not readily accommodate electric drivetrain architectures. Electric drivetrain test systems must include special considerations to account for the use of regenerative braking and electrical system components not found in conventional drivetrains, including the harmonics generated by the inverters and support for power electronics (Gavine, 2011). Development of drivetrain test benches for conventional vehicles is a well‐understood science, based upon many years of experience. However, since the electric drive train and battery components differ greatly from conventional vehicle drivetrains, experience gained from these applications is only partially transferable to electric vehicle, battery, and grid testing (Uhlenbrock, 2010)
. There are independent test stands for traction drives and batteries but none that integrate the two with grid regeneration. For battery‐grid interaction, most test labs emulate the battery using a suitable power supply, or utilize a scaled battery. Few test stands utilize full‐scale automotive battery pack, including the associated protection and thermal management systems. 2 1.2 Motivation for Grid‐Attached‐Storage Research Grid‐attached‐storage (GAS) research studies the use of energy storage that is connected to the grid and can be used to either discharge or store energy when desired. GAS research exists at two scales, based upon the size of the battery system – distributed, vehicle level, storage, such as Vehicle‐to‐Grid (V2G) ‐ and scale‐model testing of utility‐scale storage. As more electric vehicles (EVs) enter the garages of U.S. homes and businesses, utilities are faced with an increasing problem of meeting increased electrical energy demand. Utilities are also under pressure to reduce use of fossil fueled generation (Hadley & Tsvetkova, 2008)
. Consumer preferences for when to charge their EVs could challenge utilities and charging service providers by increasing grid congestion if charging occurs simultaneously with peak electricity demand. Specifically, utilities are concerned that cars that will return from the evening work commute and begin charging immediately, at a time when the demand is typically at peak levels. Solutions have been proposed to mitigate this problem include: controlled charging and using V2G to assist utilities in the provision of ancillary services, such as peak shaving (e.g. demand response), frequency regulation and renewables firming. Several studies have estimated the impact of significant penetration of plug‐in electric vehicles – whether hybrid or purely electric – on the national grid (Duvall, 2008), (Short & Denholm, 2006), (Denholm & Short, 2006)
. Pacific Northwest National Laboratory stated: “… for the United States as a whole, 84% of U.S. cars, pickup trucks and sport utility vehicles (SUVs) could be supported by the existing infrastructure, although the local percentages vary 3 by region. This has a gasoline displacement potential of 6.5 million barrels of oil equivalent per day, or 52% of the nation’s oil imports,” (Kintner‐Meyer, Schneider, & Pratt, 2006). Mike Duvall of the Electric Powertrain Research Institute, adds, “a few decades from now, there will be a 60 percent market share for plug‐in hybrids which would add 6 to 7 percent to electrical demand. The U.S. can handle most of this with existing capacity if utilities and consumers have smart charging dialed in correctly,” (Brockman, 2008). These reports highlight that significant penetrations of electric and plug‐in hybrid electric vehicles can be accommodated with existing infrastructure if charging locations and timing can be controlled. Many researchers have studied the process of utilizing electric vehicles to provide energy storage or ancillary services, known as vehicle to grid. Studies have shown that the adoption of V2G has many variations and implications (Parks, Denholm, & Markel, 2007), (Vazquez & Lukic, 2010), (Kempton, Udo, & Huber, A Test of Vehicle‐To‐Grid (V2G) for Energy Storage and Frequency Regulation in the PJM System, 2008), (Kempton, Tomić, Letendre, Brooks, & Lipman, 2001), (Kempton & Tomić, 2005), (Moura, 2006), (Turton & Moura, 2008), (Tomić & Kempton, 2007)
. More recent studies have identified challenges with the V2G model including reliability and aggregation challenges (Quinn, Zimmerle, & Bradley, 2009)
. In addition to the global and techno‐economic impacts listed above, integration of vehicles into power systems may cause local power quality and control issues that would benefit from laboratory testing of grid integration. Miller reports that, “voltage depressions and power interruptions are rapidly becoming two of the hottest topics in the field of power quality. Of particular interest is the need to supply a dependable, 4 efficient and controllable source of real and reactive power, which is available instantly to support a large (greater than 0.5 MVA) load, even if the utility connection is lost” (Miller, Zrebiec, Delmerico, & Hunt, 1996)
. Other studies have also indicated that distributed batteries may be advantageous for ancillary services and energy buffering to minimize distribution congestion (UC Davis, 2010). Distributed, vehicle‐scale battery storage provides an attractive platform for providing such services. Utilization of renewable resources such as wind, water, and solar has become a key method to reduce fossil fuel consumption for electricity generation. However, most renewable energy resources produce intermittent power, increasing the risk of voltage and frequency instability, and requiring additional power generation assets to “fill in” when renewable resources are not producing. In the case of instabilities, energy storage systems, particularly short‐term buffering or “power smoothing” resources, will become increasingly essential to grid operations. Current research focuses on modeling and simulation to understand GAS but little testing has been completed, particularly for islanded power systems or high‐speed applications of grid‐connected systems. Real results can help automotive companies, utilities, municipalities, and other intermediaries to make educated decisions on component packaging, testing, safety, and charging related to GAS. 1.3 Modes of Education for Dynamometer Laboratory As stated above in the motivation section, there exists a need for students to obtain real world, hands‐on experience to supplement traditional coursework. For vehicle studies, a lab should provide a safe environment to experience dynamometer 5 concepts, electric motors, inverters, controllers, data acquisition, high voltage safety, and experimental methods. Students can learn about torque curves, efficiency maps, and controller strategies in traditional classroom instruction, but only real world interaction with a motor test stand can translate classroom instruction into a rounded learning experience. The laboratory presented here is primarily targeted at supplementing classroom instruction, with secondary objectives as a research platform. 1.4 Modes of Research for Grid‐Attached‐Storage Several different modes of GAS research are currently being studied at national labs, universities, and in industry, at a variety of size scales. One useful method to classify GAS research is to classify based upon the operational timescale of the GAS control algorithms: Slow – operating at a time scale of minutes to hours – or fast –
operating at time scales of milliseconds to minutes. “Slow” timescale applications focus on energy‐related topics, such as arbitrage and peak shifting, which: 
Require large amounts of energy storage relative to charge/discharge rates, i.e. “low‐C” rated storage systems, 
Use relatively slow control algorithms, typically in the range of 1‐4 second control loops, 
Are primarily driven by economic criteria such as variable electricity pricing. “Fast” timescale applications focus on power‐balance topics, such as frequency regulation, which: 6 
Require relatively small amounts of energy storage relative to charge/discharge rates – i.e. “high‐C” rated storage systems, 
Use fast closed‐loop control algorithms, typically in the range of milliseconds, 
Are primarily driven by technical criteria, such as frequency or voltage stability. The classification of GAS into “fast” and “slow” categories is not clear‐cut; some applications (e.g. wind firming) could fall into either category, depending upon the objectives being optimized. Vehicle battery systems are better tuned for “fast” applications, and that is the primary target for the GAS research capability of this laboratory. It is also important to note that GAS mode of the laboratory was designed with research, not teaching, in mind. 1.5 Objectives of the eDTC The eDTC at Colorado State University will have two focuses: research and education. Corresponding to these are grid‐attached‐storage and electric vehicle dynamometer education respectively. These two focuses each have their own objectives that contribute to the key areas of study mentioned above in the motivation sections. The GAS lab is primarily focused on research, and its objectives are: 
To provide a research facility that utilizes industry‐standard components, at production scale. 7 
To provide a flexible development platform for research into GAS systems. Typical questions that could be researched in the lab include: 
Integration of GAS systems into grid systems 
Dynamic response of automotive‐type components in GAS applications 
Battery degradation and life‐time impacts of GAS operation modes 
Development of control algorithms for GAS applications The dynamometer component of the eDTC lab is primarily focused on education, and its objectives are: 
To educate students about electric vehicle components such as motors and inverters. 
To support development of laboratory hardware and coursework for vehicle power plant testing and experiments. 
To provide an environment for learning experimental methods. Typically, educational topics include: 
Dynamometer fundamentals 
Motor torque curves 
Motor‐Inverter efficiencies 
Temperature effects While each mode of operation has different objectives, the two operational modes must utilize the same components (i.e. inverters, wiring, emergency stop circuits, cooling circuits, etc.) whenever possible to reduce cost and space. 8 The next two chapters describe the GAS research mode Chapter 2 and the dynamometer mode in Chapter 3. Each chapter contains a description of the challenges, design, and results for each system. Finally, Chapter 4 presents summary results and conclusions. 9 2.
Grid‐Attached‐Storage This chapter discusses the current GAS research (Section 2.1), the design challenges (Section 2.3), design implementation (Section 2.4), and testing checkout results (Section 2.5) of the GAS portion of the eDTC lab. 2.1 Grid‐Attached‐Storage Background In order to understand the design of the GAS portion of the eDTC, this section introduces the range of GAS research topics envisioned for the laboratory. These topics include: an introduction of ancillary services (Section 2.1.1), energy arbitrage (Section 2.1.2), renewables firming (Section 2.1.3), battery second use (Section 2.1.4), and islanded power grids (Section 2.1.5). While this is not a comprehensive coverage of GAS research topics, it identifies the topics for which the lab was designed. 2.1.1 Ancillary Services Utilities must balance load and generation on a second by second basis. To maintain this balance, utilities engage in forward planning, ranging from hours to years in advance of the current period. They also maintain reserve capacity at all times, to compensate for unforeseen events such as loss of a generation unit, and to balance between load forecasts and actual load. Reserves are typically provisioned using “ancillary services” – i.e. generation services beyond simple energy production. Table 10 2‐1 (Kirby, 2004) sshows a typically set of aancillary serrvices and th
heir responsee speeds, durations, and cycle time
es. Table 2‐1: Summary of Diffferent Ancillaary Services en above fro
om Table 2‐1
1, there are m
many differeent services that utilitiess As see
must use to m
m
maintain pro
oper power ssupply each with differeent characteristics. Eyer d (Eyer sh
hows there aare many po
otentials for GAS based oon the operaational scenarios desired
& Corey, 2010)
. On
ne way to ch
hoose betwe
een the diffeerent servicees is based o
on comparingg time scales. A
At a time scaale of hours o
or longer, GA
AS units can
n be utilized for energy arbitrage – i.e
e. charging tthe battery w
when electriicity is less eexpensive an
nd then discharging w
when energyy prices are h
higher. In shhorter time fframes, milliseconds to 111 minutes, GAS units can be used for voltage and frequency regulation. Voltage and frequency control require the fastest response speeds of the types of ancillary services shown in Table 2‐1. These applications are of high importance to many researchers because they require rapid changes between charge and discharge operations that are difficult to manage and create uncertain impacts on battery life. Additionally, current research has found that the greatest near‐term economic return for grid storage is quick‐response, high‐value electric services (Brooks, 2002). In regulation and voltage control applications, batteries have several crucial strengths relative to other generation systems. Batteries have a fast dynamic response, allowing them to vary discharging (and charging) rates quickly, unlike large thermal generation units. Batteries can also transition rapidly from an off‐line status full operation, often within seconds. However, batteries are expensive, due to high up‐front capital costs and short operational lives, relative to other generation assets. The next sections discuss several applications of battery‐based GAS, both ancillary services and other applications. 2.1.2 Energy Arbitrage Infrequent but high electricity demands – commonly known as peak demands – place significant strains on grid operations. One method of reducing peak demand is to shift demand to non‐peak periods, a process commonly known as “peak shifting.” In electrical markets with time‐of‐use or real‐time pricing, peak shifting applications can be economically driven by energy arbitrage: Capturing electricity at times of low pricing, and selling it back to the utility at times with higher pricing (Sioshansi, Denholm, Jenkin, & Weiss, 12 2009)
. For example, for Pacific Gas and Electric, high prices occur during peak pricing periods, which are noon to 6:30pm and low prices occur during nighttime hours of 9:30pm to 8:30am(Pacific Gas and Electric, 2011). Energy arbitrage using existing price structures provides a method to evaluate the economics of a particular GAS technology (Eyer & Corey, 2010)
. As mentioned above, V2G is one method that has been proposed to implement energy on a large scale. The practical impacts of V2G, however, have not been fully tested. Test results could illuminate energy arbitrage performance, grid impacts, and the impact of repeated deep cycling on vehicle batteries. Vehicle owners are worried that if they participate in energy arbitrage, their car will not be charged and ready for use when needed. Since most electric vehicles for sale on the market today have an empty‐to‐full charge time of 8‐10 hours, the charging period could play a substantial role on customer acceptance (EPRI, 2011). In addition, researchers are currently modeling penetration rates, time of charge, and state of charge of battery to determine usefulness of energy arbitrage (Quinn, 2011). However, little research has utilized physical system testing to show validity of chosen control algorithms, to quantify the impact on vehicle batteries, or to help utilities and consumers make educated decisions. Utilities can also use GAS to create additional demand when system load is lower than the lowest practical operation output of its conventional generators. “Since it can take many hours to ramp up fossil fuel‐fired generation, the utility cannot simply curtail the primary generation during periods of low demand. Consequently the utility curtails renewable production instead (wind at night, solar during the day) and supplements 13 with fossil fueled backup”(Demand Energy, 2011). Alternatively, utilities could utilize GAS to save power during these periods for use during periods of high demand. 2.1.3 Renewable Energy Firming Many renewable electricity generation systems (e.g. solar or wind) typically produce power intermittently, when the resource is available. The “challenge of load balancing when the grid has a substantial portion of variable generation, and the costs associated with providing dispatchable power to back up that variable generation, raises the bar even higher [on grid control]”(Jost, 2011). One possible approach is to utilize GAS systems to convert intermittent resources into “firm” generation sources – i.e. generators that produce power output that can be predicted hours to days in advance. This application is commonly known as “firming,” and is most often applied to wind farms as “wind firming.” In this case, GAS allows a renewable power producer to establish a firm power production contract with a utility. To account for unpredictability in power production, the power producer utilizes a GAS system to ensure a consistent power output over the contract period (Walawalkar & Apt, 2008). Where large amounts of renewables are connected to weak or small grid systems, variable renewable power output can also create local frequency or voltage instabilities. Several studies have been conducted to determine the technical and economic feasibility of GAS as a method for smoothing unpredictable frequency spikes (millisecond time scale) (Bredenberg, 2011), (Sandia National Laboratories, 2010), (Goggin, 2008), (Levitan, 2010), (McGuinness & Fields, 2011)
. 14 2.1.4 Second Use Batteries Utilizing batteries after their useful life in electric vehicles, an application frequently termed “second life,” has also been a topic of recent research for GAS. The defining point for end‐of‐life for electric vehicle batteries is defined as the point when a battery has lost 20% of its original energy storage capacity or 25% of its peak power capacity (Wolkin, 2011). While this definition is somewhat arbitrary and still in flux, it recognizes that an electric vehicle depends upon its battery to provide sufficient energy to meet range expectations and peak power to meet acceleration requirements. A battery that no longer meets automotive requirements still has significant storage and power capacity … but either or both has fallen below levels acceptable to the vehicle owner. Used automotive batteries can be utilized in several applications before they are deemed unusable and need to be recycled. Applications include: back‐up power supplies, utility stationary power supplies, emergency power supplies, and renewable energy storage. In all these cases, both the initial and secondary customer can have cost savings. However, numerous barriers have been identified (Cready, Lippert, & Phil, 2003). According to Neubauer, second use of batteries is limited due to: “sensitivity to uncertain degradation rates in second use, high cost of battery refurbishment and integration, lack of market mechanisms and regulations, and the perception of used batteries” (Neubauer & Pesaran, 2010). Initial research has indicated that several alternatives exist for second use before the battery is dismantled for recycling, including grid applications such as: frequency 15 regulation, energy buffering, and energy arbitrage. Battery re‐use for grid purposes is also seen as an opportunity to reduce vehicle price through reduced recycling costs (Neubauer & Pesaran, 2010)
. Other battery cost reduction options that may drive vehicle cost down include leasing and increased standardization (Williams, 2010). Research at EPRI (Electric Power Research Institute), University of California at Davis, Sandia, and NREL (National Renewable Energy Laboratory) has made suggestions for “second life” applications, but little of this research includes realistic physical testing. Since there are many factors that determine the viability for second life batteries, more specific case studies are required to prove economic, performance, and lifetime viability of battery second use. Even tests on new batteries, such as testing performed by Xcel Energy using batteries for frequency regulation, recommended that additional testing is necessary to determine the rate of long‐term damage caused by rapid, frequent charging and discharging of the battery (Himelic & Novachek, 2010). Southern California Edison and DTE Energy also conducted real world testing of GAS applications. They used a large system constructed using battery packs built by A123Systems to charge and discharge energy at large wind power and solar power sites (Gupta, 2009). These field tests used new batteries since large‐scale packs of used batteries are currently unavailable. As a result, the performance characteristics of used automotive packs are unknown. 2.1.5 Islanded Power Grids, Expeditionary Military, or Emergency Uses The stabilization of electric microgrids utilized in emergency or expeditionary military environments represents a technically distinct application of GAS systems. 16 These microgrids are small “island” power systems, unattached to any continental‐scale grid. They are characterized by multiple, small, local generation sources servicing local loads. Expeditionary military bases face some of the highest fuel costs for electricity generation (Coomb, 2011), (Sheets, 2011). Therefore, these system operators have a strong incentive to implement renewables to displace fuel consumption. However, implementation of high‐penetration renewables on islanded grids may cause significant voltage and frequency transients due to instantaneous power imbalance between load and generation. In these applications, GAS could be implemented at a scale similar to automotive battery applications – typically 5‐100KW – to dynamically eliminate power imbalance, and thus the associated transients. This makes the use of automotive‐grade equipment an attractive source of robust, inexpensive components for small GAS systems. While this application has been modeled in simulation studies, physical testing has been rare. In addition, physical testing is required to better understand component requirements and integration into microgrid systems. 2.2 System Overview The purpose of this section is to introduce the key components of the GAS laboratory as a basis for subsequent discussion. The key components necessary for grid‐attached‐storage are: a battery pack, an inverter, an LC filter, and a transformer. Figure 2‐1 shows a simplified block diagram of the major components of the GAS system and their configuration. 17 Figure
e 2‐1: Simplifie
ed Layout of G rid‐Attached‐SStorage Equipm
ment orking left, th
he componeents Beginning at the rright hand side of Figuree 2‐1 and wo
are: he electric ve
ehicle batterry pack, whicch stores eleectricity. 1) Th
2) Th
he inverter, w
which invertts the DC volltage into grrid‐frequenccy AC when disscharging, and rectifies AC grid pow
wer into DC p
power when charging. 3) Th
he LC filter w
with a pre‐ch
harge circuit,, which filterrs out the sw
witching haarmonics cre
eated in the inverter. 4) Th
he transform
mer steps up the inverterr voltage to the grid volttage, which is 48
80VAC in this laboratory.
5) Be
etween the iinverter and the LC filterr is a voltagee measurem
ment device, the ou
utput of which is used to
o determine voltage and
d frequency output of th
he invverter. 6) Th
he voltage an
nd frequency measurem
ments are fed
d to a contro
oller, which
ad
djusts set points for the inverter andd battery maanagement ssystem (BMSS) to meet system p
performance
e goals. 188 The key component to the system is the variable‐frequency drive (VFD), also referred to as an inverter, which is equipped with a line synchronization card to allow the drive to synchronize with the grid as a line‐synchronized inverter. The system also contains protection elements, such as ground‐fault circuit interruption (GFCI) to monitor battery isolation, a battery management system (BMS), and circuit disconnects. The following section will discuss the challenges associated with these components. 2.3 Challenges to Overcome This section of the grid‐attached‐storage lab describes the challenges with designing and constructing a lab with automotive grade, full scale components. Key design issues include the compliance with standards (Section 2.3.1), differences in cooling requirements between industrial and automotive equipment (Section 2.3.2), voltage mismatch between the automotive battery pack and the grid connection (Section 2.3.3), and time delay impacts on designing a voltage/frequency measurement device (Section 2.3.4). 2.3.1 Comply with Standards While the utilization of automotive components provides unique research capabilities, these components create special concerns with grid‐attached operation. Inverter output must meet appropriate grid interconnection standards, including IEEE‐
519, thus requiring an LC filter and active front end (AFE) control. During normal operation the variable frequency drive (VFD) produces significant current harmonics, which can cause substantial impacts to other local loads or violate grid interconnection 19 standards. Current harmonics can distort the supply voltage, overload electrical distribution equipment (such as transformers) and resonate with power factor correction (Hoevenaars, 2003). The standard IEEE‐519 sets requirements for harmonic control in electrical power systems. IEEE‐519 was first introduced in 1981 to provide direction on dealing with harmonics introduced by static power converters and other nonlinear loads. Section 10.1 of IEEE‐519 says: “The recommendation described in this document attempts to reduce the harmonic effects at any point in the entire system by establishing limits on certain harmonic indices (currents and voltages) at the point of common coupling (PCC), a point of metering, or any point as long as both the utility and the consumer can either access the point for direct measurement of the harmonic indices meaningful to both or can estimate the harmonic indices at point of interference (POI) through mutually agreeable methods.” (Blooming & Carnovale, 2007). Table 2‐2 and Table 2‐3 highlight the limits recommended by IEEE‐519. 20 Table
e 2‐2: Current D
Distortion Limits for Generall Distribution SSystems (ISC an
nd IL are defineed as system short c
s
circuit and load
d size respectively) Notice
e that there are stricter limits for larrger load and weaker syystems as weell as fo
or higher ord
der current h
harmonics.
Table 2‐3: Voltagee Distortion Lim
mits As the
e use of variaable frequen
ncy drives annd other non
n‐linear load
ds has grown
n more commo
m
on, IEEE‐519 is being ado
opted by moore utilities aas a standard
d. In the particular casse of the eDTTC laboratorry, Fort Colli ns Utilities rrequires IEEEE‐519 co
ompliance fo
or all grid‐atttached devices which geenerate elecctricity, eitheer as a primaary fu
unction (e.g.. photovoltaic system) o
or as a side eeffect of usagge (e.g. regeenerative VFD). 211 2.3.2 Unusual Cooling Requirements for Industrial Labs Most industrial inverters or VFDs are air‐cooled while automotive units require liquid cooling loops. Typically, industrial inverters have sufficient space and are therefore mounted in large cabinets using forced‐air cooling. In contrast, automotive inverters are designed to be mounted in confined spaces within vehicles where there is insufficient space or access to outside air to utilize air‐cooling. Therefore, automotive inverters and motors typically utilize liquid cooling circuits. This circuit presents design issues with the construction of the lab. These additional requirements include layout, temperature control, and proper pumping performance. Therefore, an automotive‐
grade water‐cooling circuit must be implemented, utilizing a radiator and electric fan combination, similar to the set‐up in a vehicle. Given the required cooling specifications from the inverter and motor suppliers, a liquid cooling circuit was constructed. Figure 2‐2 illustrates the major system components. Figure 2‐2: Liquid Cooling System Layout 22 The cooling system is divided into two identical circuits, one for each inverter‐
motor pair, to ensure proper cooling when the units are asymmetrically loaded. Each circuit utilizes a continuous 10gpm centrifugal pump running through a flow meter, an automotive radiator with an electric fan, and an automotive closed loop thermocouple relay switch to activate the fans. The cooling circuit uses electric fans and radiators from a 1992‐1998 Honda Del Sol. These were chosen based on small size, 12VDC operation, and low cost. A decision was made to isolate thermal management from the control system to simplify control modifications for research projects. This adds a degree of separation between the control system of the lab and the cooling circuit, increasing overall system safety should the computer for the controller crash. Automotive thermocouples and relays are used to monitor the cooling circuit and activate cooling fans, without requiring any interface from the control system. 2.3.3 Voltage Mismatch The laboratory is connected to the grid at the standard industrial service voltage of 480VAC 3‐Phase. In contrast, the battery voltage typical of automotive applications is typically between 350 and 250VDC. As the state of charge (SOC) of the battery declines, the voltage of the pack decreases. The automotive VFDs utilized for the laboratory do not contain a “boost” section, typical of photovoltaic inverters. Therefore the maximum AC output voltage is constrained by the DC voltage applied to the inverter by the battery. The VFD utilized in the eDTC outputs 208VAC 3‐Phase at mid‐to‐high battery voltages. At lower battery voltages the inverter output voltage drops as low as 184 VAC. 23 This lower voltage is not a common voltage for transformers, which are designed for 208‐480VAC. Ideally, a custom‐designed transformer would be procured with the optimum turns ratio of 184:408. However, custom transformers were beyond the budget available for this laboratory. Instead, the lab utilizes standard 208:480 transformer equipped with a ‐10% connection tap on the 208 V side. Use of the ‐10% tap provides an effective turns ratio of 187:480, sufficiently close to the 184 V for operation of most of the usable range of the battery. In field applications, two methods exist to eliminate this voltage mismatch, should companies wish to reduce the size, cost, and complexity of similar systems. The first method is to specify a higher voltage battery, allowing the inverter to operate at the grid voltage over the entire voltage range of the battery. This option would require a 600 to 750VDC battery pack for direct interconnection to 480VAC. The tradeoffs to this design include costs related to the higher DC voltage and a significantly larger battery system. Alternatively, the inverter could be equipped with a boost converter to match the DC battery voltage to the desired AC voltage. Since the boost section is unneeded during normal vehicle operation, a bypass circuit or bypass mode would be required. 2.3.4 Time Delay Impact on Voltage/Frequency Measuring Device In all operation modes, accurate measurement of voltage and frequency is required in order to properly control the output of the GAS system. In addition, for stability applications, such as the small‐islanded systems described earlier, the latency of voltage and frequency measurements is particularly critical. Measurements must be completed quickly enough to support high‐speed control algorithms. The active control 24 loop can be represented by five delay components: voltage measurement, frequency calculation, controller latency, inverter response, and LC filter response. This loop represents the time required to extract information from the voltage signal, exercise the feedback control algorithm, and then execute the control signal in the inverter. Figure 2‐3 highlights these steps. Figure 2‐3: Control Loop for GAS Systems Since control algorithms will frequently change during the life of the system, all potential algorithms will rely upon the fundamental measurement capabilities designed into the system. These measurements need to be carefully designed to allow the widest range of algorithms to be tested. And according to Zimmerle, simulation and test results indicate that a key contributor to GAS system performance is the measurement delay in frequency measurements (Lutz, Zimmerle, Huff, & Bradley, 2011). 25 L2 Norm of Frequency Deviation (L2( f)/s)
Im
mpact of Mea
asurement Delay
D
on Fre
equency Con
ntrol
for Microgrid
M
Sy
ystem with 41%
4
Wind Power Penetrration
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
0.05
0.1
0
0.15
0.2
Measureme
ent Delay (s)
0.25
0.3
Figure 2
2‐4: Impact of Measuremen t Delay on Miccrogrid Perform
mance Figure
e 2‐4 illustrates typical re
esults from aa simulation
n of a small m
microgrid wiith co
onstant load
d and a wind
d power penetration of aapproximateely 41%. In th
his test, the GAS syystem is resp
ponsible for primary frequency conttrol – i.e. thee control sysstem reads syystem frequency and ad
djusts batteryy output to maintain a cconstant freq
quency set point, as illustrated in Figgure 2‐3. For this experiment, all co
ontrol system
m functions aare delay in the ffrequency m
measuremen
nt function. held constantt, except forr a variable d
em is shown as an L2 norrm of frequeency error, n
normalized to
o Performance of the syste
per‐second vaalues. Equattion 1, shown below, higghlights this.. From Figure 2‐4, m
measuremen
nt delays in the range off 0‐8 cycles ((0‐130 ms) h
have minimal impa
m
act on contro
ol system pe
erformance. Above 8 cyycles, perform
mance degraades quickly. The sensing delaay can be considered as the sum of d
delays from each of the principal com
mponents described earlier, as show n Eqn. 2: 266 Where is the
e time requirred to collecct the voltagge waveform
m, is th
he time required
d to convert the voltage waveform tto a frequen
ncy measurement, th
he controller cycle time,, is is the
e control ressponse time of the VFD//inverter, an
nd is the ffirst‐order time constantt of the outpput stages off the inverteer, which is dominated byy the first‐orrder time constant of th e LC filter. The ch
hosen autom
motive inverters exhibit fast control response tim
mes, with 10
1.5
. The outp
put LC filter exhibits a firrst‐order tim
me constant where . Utilizing a high‐spe
eed controlleer with anticcipated conttrol algorithm
ms, it iss believed th
hat the contrrol latency sh
hould not exxceed 20 mss. However, most low‐co
ost frrequency traansducers exxhibit measu
urement respponse timess of 100‐200 ms while higher‐cost trransducers, often costing $800‐20000 as an integgrated comp
ponent of a opriate for ssmall grid sysstems. To m
meet cost an
nd performan
nce power meter, are inappro
objectives, a decision was made to co
ompute freqquency direcctly from voltage wavefo
orms. This was achieved using aa zero‐crossiing algorithm
m capable off producing a frequencyy measuremen
m
t every ½ cyycle, from samples colleccted during tthe previouss half cycle. The re
esulting delaay is approximately dataa collection ttime plus computational time, or ≅ 17
for a 60 Hz system. Du
ue to measuurement noisse, the comp
puted frequeency must be filter
m
red. While th
he signal filter may be m
modified durring control aalgorithm developmentt, initial estim
mates indicaate filter cha racteristics w
will producee and effective delay of 20
60
. SSumming esttimated resp
ponse times, the system
m design will exxhibit a total control resp
ponse time oof60
277 99
,, which is within the control requirements of 0‐130ms as described above. Simulation results indicate that this response time is adequate for controlling hybrid diesel‐wind grid systems operating as small as 100 KW, although control of smaller grids may also be possible. 2.4 Design Details on the choices for key components are described in this section. Specifics of the inverter selection (Section 2.4.1) as well as its software (Section 2.4.2), the battery pack (Section 2.4.3), transformer (Section 2.4.4), LC filter (Section 2.4.5), cooling circuit (Section 2.4.6), and voltage/frequency measurement device (Section 2.4.7) are included in the following subsections. Lastly, a discussion on the designed safety (Section 2.4.8) is included. The core portion of the system was sourced from Artisan Vehicle Systems (AVS) and was delivered as an integrated system. 2.4.1 Inverter/Variable Frequency Drive Several suppliers were considered for the selection of the inverters or variable frequency drives. Selections were narrowed based on automotive component use, dual‐
purpose (GAS and dynamometer) capabilities, cost, and design complexity. Table 2‐4 describes three different package options: an Artisan Vehicle Systems (previously called Calmotors) assembly, a combination between an existing Advanced Energy (AE) power supply and AVS, and third, an entire Eaton supplied package. 28 Table 2‐4: VFD Selection Comparison Chart Package Specifications Three AVS VFDs plus LS Card AE DC Power Supply plus two AVS VFDs (1) AE Summit (3) 890CD Frame F, 57000007D, (2) Line Sync Card, AFE, 890CD Frame F, Line Pre‐Charge, LC Sync Card, AFE, Pre‐
Filter, Connection Charge, LC Filter, Kit Connection Kit Allows for Grid Connectivity for both dyno and GAS? Yes Cost $31,200 Design Complexity Simplest Partially. Dyno: use a simulated battery (power supply), GAS: use a simulated load (P Q demand) $22,800 Three Eaton VFDs (1) RGX10014AAP1 Regenerative Drive, (2) SPI140A1‐4A3N1 Inverter Drive, AFE, Pre‐Charge, DC Chopper, Sine wave Filter No. Only GAS $35,000 Missing parts (motors, PCM, and battery pack) would Requires heavy be from AVS so modifications when there would need to changing between be a lot of dyno and GAS cooperation modes between AVS and Eaton when designing the lab Based on the results from Table 2‐4, it was determined that selection number two, the combination design between AVS and AE, would best fit all the requirements of the lab except for regeneration to the grid during dyno mode. The requirement for vehicle simulation favored this implementation of the AVS solution, as the Eaton solution did not have automotive drives. The weakness to this solution is that an additional component (AE DC Power Supply) is needed. Fortunately, the eDTC lab is 29 housed within the InteGrid laboratory of the Engines and Energy Conversion Laboratory, which has volunteered an AE DC Power Supply for utilization when it’s not in use. 2.4.2 Active Front End and Line Sync Card The active front end (AFE) and line sync card (LS) were sourced from AVS because of the compatibility between the GAS and dynamometer (dyno) mode set‐up using the same automotive grade inverter. The LS allows the AFE to monitor the three‐
phase supply voltage waveform and synchronize the inverter to the grid. Once synchronized, the AFE acts as a four‐quadrant, sinusoidal, bi‐directional power supply. Since the AFE and LS card are options for the inverter, one inverter can serve two purposes. However changeovers between software, hardware, and cable connections need to be made when switching between GAS and dyno set‐up with the inverter. 2.4.3 Battery Pack The battery pack was sourced from AVS because it has the same scale and performance as full‐size automotive electric vehicle packs. This pack came assembled with a battery management system (BMS), as well as positive, negative, and pre‐charge contactors, all packaged within an air‐cooled case. The manufacturer of the cells is GBS. Specifications are listed below in Table 2‐5. 30 Table 2‐5: GBS 40Ah Batttery Cell Speccifications Several components of Table 2‐5 require further defiinition includ
ding: capacitty, ch
harging currrent, discharrging currentt and powerr density. Thee capacity of each cell iss 40Ah, which is the amoun
nt of electricc charge, whhich it can sttore. This performance is of batteries currently utilized in elecctric vehicless. The charging and faairly typical o
discharging current ratingg is given in units of C, w
which is a meeasurement of the co
ontinuous cu
urrent draw that the cell will supporrt. For exam
mple, this 40A
Ah cell with a 10C impulse d
discharge rating can disccharge at 4000A during an impulse. P
Power densitty is th
he ratio of power over w
weight, by ce
ell, and indiccates how heeavy the cellls are compaared to
o other cells producing tthe same po
ower. pack features 110 cells inn series laid out in a custom designeed The AVS battery p
enclosure to house all the
e cells, fusess, contactorss, wiring, gro
ound fault ciircuit interru
upter 311 (G
GFCI), and proper ventilaation. Since the pack is ddesigned forr lab use, thee spacing of the modules with
m
hin the pack was increassed for ease of access an
nd air circulaation using o
only one fan. Therre are built‐in air passages under thee pack to flo
ow cool air under and then over the batteries. In a pu
ure automottive applicattion, the cellls would be packed morre tightly and multiple fans would be ussed with diffferent air passages. How
wever, since the used for lab w
work, it wass deemed neecessary to h
have easy acccess to pack will be u
omponents should anything fail, nee
ed more speecific testingg, or modifications. Figurre co
2‐5 shows the
e pack and itt’s key comp
ponents. Figgure 2‐5: Detaiiled Highlightss of Battery Paack Components 322 A closser view of each cell, Figure 2‐6, reveeals the BMSS cell voltage sensors. The BMS monitorrs each cell vvoltage, currrent, and tem
mperature b
but only repo
orts the maxx vaalue of the h
highest cell tto the user. Fiigure 2‐6: Cell Monitors for EEach Cell of th
he Battery Packk 2.4.4 Transfformer The trransformer w
was sized baased on the ooutput from
m the VFD and battery du
ue to oltage mism
match as discussed in the
e challenges section. Quotes were seent out to vo
different suppliers to pro
ovide a transformer thatt was designeed for threee‐phase, 150kVA, 208V primaryy, and 480Y//277V secondary. The low
west cost biid was choseen from Je
efferson Elecctric. Specifications are sshown in Ta ble 2‐6. 333 Table 2‐6: Jefferso
on Electric 150kkVA Transform
mer Specificatiions The JJefferson Ele
ectric 150kV
VA transform
mer is a dry‐type aluminu
um‐wound NEMA TP1 sty
N
yle transform
mer, rated fo
or a 150 °C ttemperaturee rise and 22
20 °C insulation. Itt is housed in
n a ventilate
ed, floor‐mou
unt NEMA 1 enclosure. TThe part num
mber for futture re
eference is: 423‐7261‐00
00. The tapss are listed a s: 1 at 5% FC
CAN and 2 at 5% FCBN, which means
w
that the seccondary can be tapped aas low as 187.2VAC for a nominal 480 VAC primary voltaage. 2.4.5 LC Filtter An LC filter and prre‐charge cirrcuit were d esigned and
d sourced byy AVS with th
he help of Parke
er to meet th
he IEEE‐519 sstandards. TThe filter waas designed tto stabilize 344 within a few seconds at power up. The cutoff frequency was designed for less than half the inverter switching frequency (2.5 kHz). The filter operates as an LCL filter, with the transformer providing the secondary inductance. Parker was chosen to design the LC filter to maintain capability with the inverter. The final components were modeled in SimPower Systems to determine the response time of a pre‐charge filter. However, since Jefferson did not provide sufficiently detailed design information, it was only possible to create a basic model. The model indicated that the pre‐charge resistors needed to be greater than 0.25 Ω in order to limit the in‐rush current to less than three times the nominal current. Due to concerns with the model, it was decided to utilize prior experience by AVS to size the charging resistors. As a result, AVS sized the resistors at 45 Ω. Figure 2‐7 presents the circuit schematic of the LC filter cabinet packaging, including the pre‐charge circuit. 35 Figu
ure 2‐7: LC Filte
er Cabinet wit h Pre‐Charge C
Circuit Schemaatic 2.4.6 Coolin
ng Circuit The liq
quid coolingg circuit was implementeed based on the cooling requiremen
nts provided from
m Artisan Ve
ehicle System
ms and addreesses the ch
hallenges preesented in 366 Se
ection 2.3.2. Photos in FFigure 2‐8 an
nd Figure 2‐99 illustrate kkey componeents of the co
ooling circuit. Figure 2‐8: C
Cooling Pumpss with Primingg Valves, Flow Meters, and FFoot Valves Figure 2‐9: Coo
F
oling Circuit Fans, Radiators, and Reservoirr 2.4.7 Voltagge and Frequency Senso
or As desscribed prevviously, GAS control algoorithms requ
uire fast meaasurement o
of vo
oltage and frequency. TThe sensor m
must supportt 480VAC opeeration, fit w
within the 377 budget, and produce fast measurement time (less than 0.13s; see section 2.3.4). Table 2‐7 highlights the results of the various sensors considered for selection. Table 2‐7: Voltage/Frequency Sensor Comparison AC Sensor Voltage Response Time Price Output Range Ohio Semitronics Inc. DMT‐
±20mA, 100‐636V <100ms $1,405 1024 RS‐485 4‐20mA, Ohio Semitronics DSP‐008 600V <100ms $1,025 RS‐422 Elkor WattsOn‐1200‐MSCT1 480V 500ms Power True Systems Corp. 2V‐33‐AD‐6‐A5‐D2 and 2F‐33‐
0‐500V Not 0‐10V or Available 4‐20mA A/D conversion not included Custom Built Voltage Divider As fast as LabView cRio Data Acquisition RS‐485 <400ms AD‐6‐A5‐D2 and National Instruments $376 0‐750V program is $480 ±4.8V designed As seen above, the best sensors with the lowest response time (Ohio Semitronics) are also the most expensive solutions. Therefore a custom‐built voltage divider with NI DAQ solution was chosen. The sensor was built using a precision decade resistor. The 1776‐C54 voltage divider by Caddock Electronics is shown in Figure 2‐10. 38 The divider provides 1MΩ
Ω of resistan
nce across thhe divider wiith taps at 1K, 9K, 90K, aand 900K Ω. Figure 2‐10
0: Caddock Ele ctronics Voltage Divider The output was atttached to p
pins 3 and 5 so that the D
DAQ would ssee ±4.8V frrom th
he voltage and frequenccy voltage divider (a 10kkΩ total resisstance). It is noted that aa caapacitor can
n be added laater for filtering if the siggnal is too n
noisy. 2.4.8 Uniqu
ue Safety Re
equirementss Uniqu
ue safety req
quirements w
were encounntered due tto the naturee, size and vo
oltage levelss of the equiipment. Nott only is the system diffeerent than trraditional in
ndustrial app
plications be
ecause it utilizes automootive equipm
ment, but it aalso function
ns in an educational atmosphe
ere. In this environment the system has two diffferent users: re
esearchers, w
who need fle
exibility, and
d students w
who may nott be experien
nced with 399 electrically and mechanically dangerous equipment. Both of these users require different safety measures. Relative to traditional safety standards, the system needs to have ground fault detection, multiple E‐stop switch locations, and thermal management. The insulation or ground fault detection device is discussed in the next section. Emergency stop buttons are located throughout the test center in three different areas to ensure that the user is always within ten feet of an E‐Stop. In typical industrial environments E‐stop switches are included and installed in covers and doors of the purchased system but since the components were sourced from an automotive supplier, these switches were added during the lab design and construction. More specifics of the E‐stop system are detailed in section 3.4.10. In addition to this safety, light towers were installed to provide visual verification if a breaker is closed and the system is live. This will visually notify users not to open breakers, control boxes, make electrical connections or perform repairs when the system is energized. The battery management system provided by Artisan monitors each cell performance to determine battery safety and health. These values are communicated to the Powertrain Control Module graphical user interface (PCM GUI). The PCM GUI monitors the battery SOC and prevents operation if the battery SOC is not within the defined operating range. For example, if the battery were fully charged when the GAS controller commands a charging operation, which could over‐charge the battery, the PCM would cancel the command in order to preserve battery health and proper use. The PCM GUI software has limits defined for cell over/under voltage, over/under temperature, and charge/discharge rate. 40 2.4.8.1 Insulation Moniitoring Devicce Batterry subsystem
ms present u
unique safetyy challengess, particularly guarding aggainst short circuits betw
ween batterry supply linees and groun
nd. The systtem is equip
pped with a ground
w
d fault circuiit interrupter (GFCI), whhich engagess when the system is nott operational. If for any re
eason the battery becom
mes grounded, the GFCI opens the baattery contaacts and sounnds approprriate alarms.. Note that tthe disconnects/o
GFCI system i
G
is operational only when
n the batteryy is not conn
nected to an
nd operatingg as part of the syystem. The Bender IR155
5 was chosen
n because it provides a cclear “ok” – safe or “nott ok” ut. The IR155
5 monitors th
he insulationn resistance between th
he insulated and unsafe outpu
active high vo
oltage lines o
of battery an
nd the refereence (chassiis) ground. IIts design for co
ompact size,, temperature, and vibraation makess it ideally su
uited for hyb
brid vehicle applications. Figure 2‐11 (image from
m: http://ww
ww.iso‐f1.de/downloadss/iso‐F1‐IR15
55‐
32xx‐electric‐‐vehicles_DB
B_en_20101
1202.pdf) shoows the Ben
nder IR155 G
GFCI used in the laab. Figgure 2‐11: Ben
nder IR155 Gro
ound Fault Circcuit Interrupteer 411 2.4.8.2 Insulated Conneectors Since the lab is de
esigned to op
perate in tw
wo independent modes, GAS and dyn
no, a saafe way to sw
witch betwe
een the two was requireed. To convert to GAS mode, the invverter co
onnections n
need to be u
unplugged frrom the mottor and attacched to the LC filter. Theese in
nterconnectiions would b
be difficult to
o produce u sing interloccked breakers and switcches, opening the p
possibility th
hat the syste
em could be started in an invalid con
nfiguration. nstead, the cconnections are perform
med using coolored cablin
ng and shield
ded quick In
disconnect caam‐lock conn
nectors from
m Marinco E lectrical. Theese connectors, which aare wer and exp
peditionary m
military operations, safeely handle th
he also utilized ffor event pow
nd allow safee interconneect while minimizing thee risk high current aand voltage demands an
of a short circcuit. Figgure 2‐12: Cam
m Connectors b
by Marinco. Clo
ockwise: Femaale Cam Conneector, Male Caam Connector, and Demo
onstration of I nstallation, Paanel Mounts 422 Figure 2‐12 (images from: http://www.beckelectric.com/store/pc/Marinco‐
Information‐c266.htm) illustrates the different cam connectors. They have a male or female plug to prevent unintended connections. The system was designed such that the user only matches the proper colors. The polarity and color of the plug were designed so that only one arrangement is possible. Color‐coding was selected to minimize the chance of inadvertently connecting incompatible power leads. DC leads are color coded as black (V‐) and red (V+). The AC connections would normally use the industrial standard of red‐white‐blue for 208 VAC. However, this would create the potential to interconnect a red DC line with phase A of the 208 AC lines. Therefore, all AC lines are color‐coded using the 480 VAC standard of brown‐orange‐yellow for the three phases. With the use of these insulated connectors, users of the lab can quickly, safely, and with confidence, change between the two lab set‐ups. 2.5 Check Out and Verification This section describes overall lab construction (Section 2.5.1), a step‐by‐step overview of the system operation (Section 2.5.2), and results of the system verification tests (Section 2.5.3). 2.5.1 Finalized System Construction A 3D CAD model was utilized to layout the lab. Figure 2‐13 illustrates this model with call‐outs to the specific components of the lab. 43 Figure 2‐13: Overview
w of Compone
ents and Their Locations in 33D CAD Model The C
CAD model illlustrates key componennts; conduitss and cablingg are not sho
own. During constr
D
ruction each
h conduit waas sized accoording to thee required w
wire size and fill raatio as dictatted by the lo
ocal electricaal code. Figure
e 2‐14 showss the comple
eted construuction of thee GAS portio
on of the eDTTC laab with call‐o
outs labelingg the major ccomponentss. 444 Figure 2‐14: Actual Layo
out of GAS Com
mponents Figure
e 2‐15 showss the specificc details of tthe components installeed in the LC ffilter caabinet. The aattenuator is used as fee
edback for tthe inverter to communiicate with th
he filter and pre‐charge circuit. Both the
e pre‐chargee and the maain circuit co
ontactors aree housed in the
e one packagge labeled co
ontactor. 455 Figure 2‐15: Actual Layyout LC Filter C
Cabinet 2.5.2 System
m Operation
n The syystem workss by booting up the conttrol computeer and opening the PCM
M GUI so
oftware and the Parker SSSD Drive co
ontrol software, called D
DSE Lite. Theen the user m
must co
onfirm CAN communication is estab
blished with the softwarre to ensure proper systeem ch
heckout. Once these ste
eps are prefo
ormed the u ser must peerform safetyy checks to ensure that: n
no cables are improperlyy color matcched, everytthing is prop
perly connected, all shields and
d covers are
e closed, all EE‐stop switchhes are in th
he up (releassed) position
n. It iss recommended that an experienced
d supervisorr confirm thee correct opeerational staate before approving the testting procedu
ure. Then tthe InteGrid
d switchgear main servic e disconneccts need to b
be switched on eaker can bee switched to
o the ON po
osition. The LLC (cclosed) and tthe 480VAC ““I‐Beam” bre
filter pre‐charge switch M
MUST be in tthe off positiion before cclosing the LC
C filter breakker 466 to
o prevent the LC filter circuit from arcing. The fi lter cabinet fans and thee radiator faans sh
hould be acttivated at this time. The pump for thhe GAS circu
uit must also be plugged in. At this point t
A
the breaker to the LC filtter can be sw
witched to tthe ON posittion. The useer must ensure m
that the GASS inverter is communicaating properlly and seeing an AC bus vo
oltage of 187VAC. When
n the proper voltage is oobserved, thee LC pre‐chaarge contactors caan be switch
hed ON to be
egin the pre‐charge. After a few secconds, the m
main contacto
or will automati
w
cally close w
with a loud “thud.” Afterr system cheeckout, the b
battery key‐o
on sw
witch may be turned to begin the baattery pre‐chharge and fu
ull connectio
on. The final step in
nvolves the u
user flippingg the “ENABLLE” switch foor the GAS in
nverter on th
he desk conttrol box (shown in
n Figure 2‐16
6). Figure 2‐16: Desk Control Box with Swittch Callouts Usingg the DSE Lite software, the user cann define currrent control mode, moto
or in
nduction leakage, and cu
urrent demaand from thee battery. Byy defining the current demand values the batte
ery will chargge or discharrge accordin
ngly. The useer can monittor 477 vaarious param
meters including: real an
nd reactive ppower, poweer angle, pow
wer factor, actual current, DC voltage, real and rreactive currrent feedbacck, and synch
hronization frrequency. An
n example o
of these paraameters is shhown in Figu
ure 2‐17, below. Figurre 2‐17: Displaay Parameters in DSE Lite witth Arbitrary V
Values for Display Purposes O
Only 2.5.3 Testin
ng Results Operaating the sysstem to charge and dischharge of the battery from
m the grid performed a final system
m checkout. O
Only manuall current con
ntrol was utiilized. Duringg manual contr
m
rol, the inverrter current was set to ++10A and ‐100A for conseecutive ten se
econd period
ds. Figure 2‐‐18 is a scree
enshot show
wing the systtem synced tto the grid aand th
he battery discharging and then chaarging. In thee background is the DES Lite inverter so
oftware show
wing grid synchronizatio
on, true currrent control mode, and 10A current limit. In the fo
oreground iss the PCM GUI showing battery volttage in the lo
ower left corrner. Notice the di
N
p towards th
he end of the plot. This ddesignates tthe dischargee period of tthe te
esting. A mo
ore detailed vview of even
nt is shown iin Figure 2‐119. 488 Figure
e 2‐18: Screenshot of DSE Litte in Backgrou
und with PCM GUI BMS Mon
nitor in Foregro
ound Figurre 2‐19: PCM G
GUI BMS Plot o
of Pack Currentt and Voltage vs. Time Plot ((left) and Maxx and Min Cell V
Voltage vs. Tim
me Plot (right)
Duringg this testingg the scope function of D
DSE Lite wass utilized to monitor thee re
eactive power of the sysstem as well as the voltaage comman
nd along with
h actual volttage (sshown in Figgure 2‐20). 499 Figure 2‐20: Voltaage and VAR vvs. Time Plot fo
or DC Voltage Command and
d Real as well as Reactive Pow
wer during thee Discharge Eveent This syystem checkkout verified
d that all pow
wer compon
nents were o
operational aand fu
unctioning as intended. However, ad
dditional tunning will be rrequired forr future research applications. Several issues need to b
be noted. Reeferring to Fiigure 2‐20, tthe reactive power fluctuaates at the sstart of the test; this is d ue to sub‐op
ptimum softtware settinggs and poor DC voltage meaasurement at the batteryy. A slow ram
mp rate is allso apparentt at th
he end of the discharge test, when it takes seveeral seconds for the reacctive power tto se
ettle at the ccommanded
d level. It is ccurrently beelieved that tthese issues can be reso
olved viia software u
updates and
d further systtem configuuration testin
ng. AVS has committed tto providing technical assisttance with th
his fine‐tuninng during th
he next testin
ng session.
500 3.
Electric Vehicle Component Education through a Dynamometer This chapter focuses on the dynamometer portion of the eDTC. Sections include an introduction of to dynamometers (Section 3.1), an overview of the system (Section 3.2), the challenges that were presented with the system (Section 3.3), the design selection of each of the key components (Section 3.4), a system check‐out (Section 3.5) and course work example (Section 3.6). 3.1 Dynamometer Introduction A dynamometer (hereafter referred to as dyno) is a device used as a brake to test rotary machines. Dynamometers are utilized to test internal combustion engines, motors, generators and related machines. A dyno consists of a shaft coupling and a controllable loading mechanism, typically capable of independently controlling either torque or speed of the unit under test (UUT). There are many different kinds of dyno loading mechanisms, including: water brakes, eddy current brakes, electrical generators, fan, hydraulic resistance loads, and controllable mechanical friction loads. Each type has its own advantages and disadvantages, which needed to be considered based on the requirements of the lab. Since the eDTC lab has very tight space requirements, this limited many technologies. The electrical generator dyno type was pursued since it would be the simplest to implement and most compact. Using an electrical motor as the brake or absorber means that electrical power recovered from the dyno could be reused 51 by the system, thus requiring an external power to only provide power for the losses within the system. 3.1.1 Various Methods of Motor Testing To Date Many industrial and educational internal combustion engine (ICE) stands exist around the world, but their design is very different compared to an automotive electric motor test stand. The existing test benches have liquid cooling infrastructure, exhaust venting, emissions measurement equipment, fuel storage, plumbing, and safety precautions, as well as different sensors (O2, timing, thermocouples, pressure, mass air flow rates, etc.). However, only a few of these components and testing knowledge can be carried over to electric motor and GAS test stands. ICE test stands do not have inverters, batteries, or traction DC power supplies. In addition to the differences, some dynos (i.e. water brake dynos) do not have generators that can be used for regenerating power back on to the local utility grid. In general, test operators would prefer to sell power back to the grid. However, they are constrained by the utilities interconnection regulations and additional expensive infrastructure. For this reason, most dyno test stands today simply turn the used energy into heat and dissipate this energy during testing. This gives operators more flexibility at the expense of continuous power input. In the eDTC lab, it was desired to not use the dyno to produce waste heat but rather minimize consumption by designing a regenerative solution. While the dyno set‐
up of this lab doesn’t have grid regeneration, it does feature a regenerative loop that minimizes power consumption. This regenerative concept was selected to minimize power input and to reduce cost of components. 52 3.1.2 Uniqueness of this Set‐Up The CSU eDTC test lab is unique because the majority of current test labs support only ICE testing. Few universities have full‐scale traction motor testing, making it difficult for students to acquire hands‐on experience with electric vehicle components and for researchers to perform physical testing. Many vehicle manufacturers, component designers, and even national labs have electric motor testing facilities or hybrid dynamometers, but few have an educational laboratory utilizing components of this scale. The unique contribution of the eDTC laboratory is providing an educational facility using full‐scale automotive components in a flexible system configuration. 3.2 High‐level System Overview of the Dynamometer Lab This section describes the architecture of the dyno laboratory at a high level, including the motor to test, the absorber motor, inverter drives for each motor, power supply, and torque sensing. 53 Figure 3‐1: Simplified Diagram of Dyno Set‐Up Above, Figure 3‐1, shows selected components and how they are interconnected. Starting from the top of the diagram: 
The DC power supply supplies make‐up power to the system, i.e. the losses due to friction and impedance in the drives, motors and drivetrain. The DC power supply provides closed‐loop control of the DC bus voltage. 
The DC bus provides a common DC bus for both the driving and absorbing motor‐inverter pairs. The “driving motor” is the unit under test (UUT), while the “absorbing motor” loads the UUT. 
The UUT inverter (890 CD is the inverter model number) provides variable frequency and voltage, 3‐phase, AC power to the UUT motor, which converts electrical power into mechanical rotational power. 54 Control of the UUT inverter can be modified to support education or research needs. Either speed or torque control are possible. 
The torque sensor monitors both the torque and speed of the UUT. 
The absorber motor loads the UUT, generating variable‐frequency, 3‐
phase, AC, power to its inverter. 
The absorber inverter (also inverter model number 890 CD) absorbs power from the absorber motor and drives that power back onto the DC bus by adjusting its DC output voltage. Generally, the absorber will be controlled using the opposite parameter of the UUT inverter – speed if the UUT is controlling torque, and vice versa. The only major component not shown in the figure is the power train control module (PCM), which facilitates communication between the control computer and inverters. 3.3 Design Challenges This section elaborates on the challenges in the design and construction of the dyno portion of the lab, and describes the selection of key components. 3.3.1 Cogging and Other Vibration Effects Under certain high‐load conditions, traction motors may exhibit significant cogging. In traction‐drive test mode this cogging can excite torsional vibration modes in the drivetrain between the two motors. These vibrations may be amplified if the compliance in the couplings and the torque transducer create natural vibration modes 55 near the excitation frequency. The top design speed of the system is 8,000 RPM. Given that the motors have eight poles, the fundamental cogging frequency is 32,000 RPM or 533 Hz. To avoid resonance, an ideal solution would place the first vibration mode above 1066 Hz. Given that the mass of the UUT and absorbing motors is largely set by their size, the frequency requirement primarily impacts the selection of the torque sensor and shaft couplings. 3.3.2 Sizing of DC Power Supply The lab is not designed to utilize the battery as a motor power supply. Therefore a DC power supply was utilized to power the inverters during testing. Each motor is rated at 70kW. Therefore, if no regeneration is utilized, the lab must provide up to 70 kW to the UUT motor (plus losses), and must dissipate up to 70 kW generated by the absorber motor. However, the power supply demand can be greatly reduced if the power generated by the absorber is reused within the dyno, and the need for a 70 kW load bank can be entirely removed. In this architecture, only “make‐up” power will be required to account for friction and impedance losses within the two motor/inverters. This regenerative set‐up is achieved by placing both inverters on a shared DC bus. Therefore, the size of the make‐up supply depends on the losses in the system. To determine these losses, the worst efficiencies were used to calculate the maximum make‐up power required for different power and speed situations. Figure 3‐2 shows how the system was modeled for calculating power consumption based on the different efficiencies of the different components. 56 Figure 3‐2:: Schematic fo r Efficiency Caalculations The effficiency at e
each point w
was calculateed by addingg the losses aat each poin
nt of syystem using equation 3: where P
w
er needed att the motor shafts and eepsilon (ε) d
denotes Desired
d is the powe
effficiency for each compo
onent noted
d in subscriptt. Even thou
ugh the moto
ors are identtical, AVS indicated
A
d that the effficiency of the absorberr is better th
han the unit under test because the m
motors operrate better in
n regeneratiion. Howeveer, in order tto determinee the worst case sc
w
cenario, the better efficiencies of thee motor in regeneration
n mode weree neglected and
d the lower efficiencies (i.e. when thhe motor op
perates in po
ositive torqu
ue) were used. U
w
sing the equ
uation and known efficieencies, Tablee 3‐1 was co
ompiled. 577 Table 3‐‐1: Required M
Make‐Up Poweer by the Poweer Supply Calcu
ulations Table 3‐1 shows that an estim
mated 19kW is needed in
n the worst ccase. With these estimations aand a 25% faactor of safetty, it was deetermined th
hat an existin
ng 25kW pow
wer su
upply would
d provide suffficient make
e‐up power. Re‐using th
he existing 24
4 kW powerr su
upply reduce
ed the cost o
of the system
m by $4,1500 – the cost o
of an additio
onal 50 kW d
drive to
o provide maakeup powe
er. 588 3.4 Design
n of this Equ
uipment This secction describ
bes the key ccomponentss incorporateed in the dyno lab includ
ding: electric moto
ors (Section 3
3.4.1), invertters (Sectionn 3.4.2), DC p
power supply (Section nt sensors (Section 3.4.4
4), torque sennsor and cou
3.4.3), curren
uplers (Sectiion 3.4.5), mounting box
m
x for the mo
otors, couple
ers, and torqque sensor (SSection 3.4.6
6), powertraain co
ontrol modu
ule (Section 3
3.4.7), contrrol compute r (Section 3..4.8), data accquisition syystem (Section 3.4.9), an
nd emergency stop circuuit (Section 33.4.10). In eaach section, re
equirementss are discusssed, followed
d by the chooice of comp
ponents to m
meet the re
equirementss. Figure 3‐3 shows these major com
mponents an
nd their integgration in a more detailed view
w. Due to margin constraaints, the sizee of the imaage is reduceed thus makiing th
he lines difficult to see (ffor the full sscale image pplease refer to Appendixx A: Figure 0
0‐9). Figure 3‐3: Scchematic for eD
DTC Dynamom
meter Set‐Up 599 Starting from the left hand side of the figure: 
The service disconnect provides power to the Summit™ DC Power Supply. 
A diode prevents power flowing in reverse direction from the DC bus to the DC power supply. 
Both inverters are connected to the DC bus using Marinco insulated power connectors. 
Current sensors are shown within the DC bus bar area to monitor the current flow to each inverter from the DC bus. 3.4.1 Electric Motors For this lab, both a UUT motor and an absorber motor (acting primarily as a generator) are needed. Since the lab will function as a test bed for any automotive‐
based electric motor, the choice of motors is not particularly crucial. Motors were chosen for the dynamometer lab from Artisan Vehicle Systems because the automotive‐
style motors could be procured at below market cost and be pre‐package to work together with AVS inverters and powertrain control module (PCM), as well as an open‐
source control system. The chosen motor pair is the Parker GP150WC motors packaged and sold by AVS. They are surface mounted, three‐phase, permanent magnet synchronous motors with 8‐poles, 70kW peak output, 8000rpm maximum speed, and 138Nm maximum torque. Figure 3‐4, Figure 3‐5, and Figure 3‐6, courtesy of AVS, illustrate the torque, power, and efficiency curves of the selected motors. The high performance to weight ratio and compact packaging of these motors is typical of automotive motors. 60 Figure 3‐4: Output Torqu
ue of GP150WCC Motor from Artisan Vehiclle Systems 611 Figure 3‐5:: Output Powe
er of GP150WCC Motor from A
Artisan Vehiclee Systems Figure 3‐6: Efficiency Curvve of GP150W
WC Motor from Artisan Vehiccle Systems 622 3.4.2 Variable Frequency Drive Variable frequency drives (VFD) or inverters are needed in AC synchronous motor applications to provide speed control. The VFDs match the 3‐phase power required by the motors to the DC power available on the DC bus. Improper frequency selection could potentially lead to significant motor slip causing stalling or damage to the motor or VFD. The choice of the VFDs was constrained by requirements for the GAS operation mode, described in Section 2.4.1. While the UUT VFD is not utilized in GAS mode, the absorber VFD is reused as the line inverter for GAS mode. To convert from GAS to dyno mode, the user must disconnect the battery and LC filter from the DC bus, and connect the UUT VFD to the DC bus. Additionally, the user must replace the “line synchronization” (LS) with the resolver card necessary for dyno speed control, and the data harness must be swapped appropriately. Detailed instructions of this procedure can be found in Appendix E. Finally, the software on the inverter must be re‐
programmed to support dyno mode. 3.4.3 DC Power Supply The eDTC utilizes an existing 25kW DC power supply from the InteGrid Laboratory, a Summit™ 57000007D from Advanced Energy Incorporated. This supply provides floating DC output, which means it is insensitive to the direction of DC current flow when operational. However, in emergency shutdown events, the supply must be protected from a potential current surge caused by discharge of output capacitance of the VFDs. To provide this protection, a Vishay SD200N20PC diode was installed in the 63 DC power line
D
es to preven
nt current baackflow. Thee diode supp
ports a 200A
A forward current with a 1.4V d
w
rop, and hass reverse voltage protecction to 20000V. Figure 3‐‐7 shows thee in
nline installation of the d
diode on the
e DC positivee wire in thee DC disconnect enclosurre. Figu
ure 3‐7: DC Discconnect Intern
nals Showing D
Diode Installattion Currently, there iss no capacitaance across the DC bus, other than the output caapacitance o
of the VFDs. This producces a very “sstiff” DC buss, which mayy excite high frrequency mo
odes or conttrol instabilitties. A capa citor may bee required to
o damp thesse modes if issu
m
es are obserrved; no estiimation of thhe capacitorr requiremen
nts was co
ompleted du
uring this wo
ork. 644 3.4.4 Current Sensor To measure the VFD power, it is necessary to measure both the DC bus voltage and the current on each VFD connection. Voltage measurement was described in the GAS description, above (Section 2.4.7). Current measurement is described below. The three most common types of current sensor are: shunt, current transformer, and Hall‐effect sensors. Current transformers are only suitable for AC current measurements. Current shunts are often susceptible to noise, may have heating issues at high current levels, and often require galvanic isolation to prevent floating DC‐mode signals for damaging data acquisition equipment. Modern Hall‐effect sensors, however, gain noise immunity by using built‐in amplification, and, because they are indirect sensors, provide both galvanic isolation and virtually no heating at high current levels. “Hall effect sensors are broadly deployed because of their dc current sensing capability, galvanomagnetic isolation, and zero circuit insertion loss” (Zhou, 2003). Within the Hall‐effect measurement type there are two different groups based on their measurement control loop. These two groups are defined as open or closed loop sensors. Open loop Hall‐effect current sensor designs internalize the current carrying conductor, allowing the manufacturer to optimize the current sensor package’s size and thermal characteristics. Measurement of open loop sensors usually has a more limited range of linearity and cannot compensate for offset and residual field errors. The closed loop Hall‐effect sensors, on the other hand, act as transformers that use a reference voltage, which allows for higher accuracy than open loop control sensors. The disadvantages to closed loop sensors include larger size and higher cost. The closed loop 65 sensor is typically more accurate because the output of the Hall‐effect element is connected to a secondary coil, which generates a field to cancel the primary field’s current. The amount of cancelling is proportional to the coil ratio between the two (Dickinson & Milano, 2002)
. Another consideration is the measurement signal. All Hall‐effect sensors need an amplifier to increase the signal, but some closed loop sensors are designed to work with a data acquisition system that measure the delta voltage and interprets this as current flow. This allows for reduced component count, cost, and complexity. Based on the above knowledge, the requirements for the current sensor selection were: 425A peak measuring ability, sensor size must be smaller than the allotted space in the junction box (6”x2”x6”), closed loop for accuracy, and lowest cost relative to other closed loop options. The Honeywell CSNS300M current sensor was chosen because it has +‐12VDC reference voltage requirement, 600A max measurement, closed loop control, and ease of panel mounting. Figure 3‐8 shows the final current sensor selection installed in the high voltage DC bus bar cabinet. 66 Figurre 3‐8: DC Bus Cabinet with CCurrent Sensors Shown on R
Right The se
ensor needs measureme
ent resistorss to tune thee sensor’s ou
utput. The measuremen
m
t resistance was chosen
n as 51 Ohmss because th
he range of tthe suggesteed re
esistance waas 5 to 95 Oh
hm at 200A o
or 5 to 50 O hm at 300A max. This su
uggestion is also att a rated ±15
5VDC, where
eas the lab w
will be using ±12VDC. Du
uring initial sset‐up, if thee measured val
m
lues have po
oor resolutio
on, a differennt resistor w
will be used. 677 3.4.5 Torque Sensor and Coupler Selection 3.4.5.1 Objectives The eDTC needed to purchase a torque sensor that could be used in an electric motor dynamometer to measure the torque in the system between the motor under test and the absorber motor. The torque sensor requirements included: price, torsional stiffness, sampling rate, non‐contact measurement, operational speed of 8000rpm, and capable of withstanding 138Nm of torque. Due to the significant effect that cogging and vibration can have on measurements, finding the best sensor depended on: max measuring speed, max torque limit, cost, torsional stiffness, and moment of inertia. The following sub‐section describes the method used to make the selection. 3.4.5.2 Method Several torque sensors were investigated as valid sensors for the eDTC dynamometer application. These were tabulated below (Table 3‐2) based upon the selection parameters. In total, eleven sensors were considered for evaluation. In order to narrow down the results, sensors over $5,500, rotational speeds below 10,000 rpm and sampling rates slower than 1kHz were ignored. 68 Table 3‐2: Compilatiion of Torque Sensors Considered for Seleection Notice
e the special exceptions to the Binsffeld sensor. There is no torsional sttiffness (Kc) and no mass moment o
of inertia (J) vvalues for th
he Binsfeld ssensor becau
use it utilizes inductive strain gauge sensorrs built into aa collar that is rigidly atttached to the sh
haft. Many o
other sensorrs have no Kcc or J valuess simply becaause the com
mpany didn’t re
elease any o
of this data. 699 In ord
der to properrly determin
ne the best ccoupler for the system, tthe followingg parameters w
were used in the selectio
on choices: ttorsional sprring rate, clamping quality (sself‐centerin
ng and adequ
uate strength), speed raange, balancing quality, aalignment caapability, an
nd price. The
e selection ch
hoices are taabulated in TTable 3‐3. D
During torque se
ensor researrch, it was re
ecommende
ed that the c oupling should have at least three ttimes th
he torsional stiffness of the sensor. TThis finding helped narrrow down th
he selection. All th
he selected ccouplings be
elow complyy with this reecommendation. Table
e 3‐3: Couplingg Selection Values In order to
o further narrrow down tthe selection
n, modeling of different se
ensor‐couple
er combinattions was completed in oorder to determine the ssystem natu
ural frrequency. Du
uring this invvestigation, different meethods for d
defining the model weree determined. Figure 3‐9 highlights this model withh definitionss of each of the compon
nents. Note that the
N
ere are two m
methods of defining thee model baseed on the plaacement of tthe 700 to
orsional stifffness of the torque sensor shaft. Thee red mass n
number 4 is the definitio
on under investigation. Figgure 3‐9: Model Definitions Used to Deterrmine System N
Natural Frequency with Masss Numbers Labe
N
led above Diffferent Compon
nents After modeling ussing both me
ethods it wa s found thatt defining th
he system as first sh
hown in Figu
ure 3‐9 make
es most sensse physicallyy; therefore Caveat 2 is llikely less re
epresentativve of the phyysical system
m. Figure 3‐110 illustratess the mode shapes of th
he syystem. More
e specific infformation on
n the model ing and detaailed plot ressults can be fo
ound in Appe
endix D. 711 Figure 3‐10: Am
mplitude at eacch Different Maass Number fo
or the Three Different Modees 3.4.5.3 Resu
ults The re
esults of the frequency ccalculations provided intteresting cost vs. As the plot b
below (Figurre 3‐11) dem
monstrates, aalmost all the performance trade‐offs. A
se
ensors have a similar fre
equency rangge (330‐415 Hz) for the llowest vibration mode exxcept the Binsfeld (600H
Hz), despite substantial vvariation in price. The plot also show
ws in a red dashed line, the coggging freque
ency of the m
motor at maaximum speeed. Notice th
hat all but one se
ensor‐couple
er combinatiion has a nattural frequency lower th
han the coggging m must transsition througgh a natural frequency frrequency, indicating that the system
before reachiing maximum
m motor spe
eed. 722 Fiigure 3‐11: Vissual Representtation of Senorr and Couplingg Cost vs. Natu
ural Frequencyy Comparison w
with a Dashed
d Red Line Representing the Motor Naturaal Frequency 3.4.5.4 Torq
que Sensor SSelection The se
election of the torque se
ensor and cooupler comb
bination cam
me down to aa decision of prrice over perrformance. FFigure 3‐11 sshows that aall the options except th
he e the same n
natural frequ
uency perforrmance while having varrying price. TThe Binsfeld have
Binsfeld soluttion exceede
ed the budge
et, thereforee it was deciided to purchase the low
west ost solution:: the Sensor Tech RWT320 with BK22 couplings. FFigure 3‐12 aand Figure 3
3‐13 co
sh
how the sele
ected compo
onents. 733 Figure 3‐12: R+W Am
merica BK2 Bellows Couplingg with Clampin
ng Hub nsor Technolo
ogies RWT320 TTorque Sensorr Figure 3‐13: Sen
The Se
ensor Techn
nologies torq
que sensor uutilizes a uniq
que measuriing techniqu
ue based on meaasuring the rresonant fre
equency change of Surfaace Acoustic Waves (SAW
W). It does this in a non‐contacct manner when strain iss applied to a shaft insid
de the sensor. es a deformaation of the quartz substtrate of the SAW device, The applied torque cause
which in turn
w
causes a ch
hange in its rresonant freqquency. Thee frequency ssignal measu
uring th
he resonant frequency iss coupled viaa a non‐conttact RF rotatting couple ffrom the shaaft to a fixed pick‐u
up. It is the analysis of th
he differencee in resonant frequenciees between tthe wo SAW devvices, with ellectronic pro
ocessing andd calibration, that gives aa precise tw
in
ndication of the torque ttransmitted by the shaftt. SAW devicces have a hiigh immunitty to magnetic field
m
ds allowing ttheir use in motors, for example, wh
here other aanalog 744 technologies are susceptible to electronic interference and are not suitable (Sensor Technology, 2011)
. Since no sufficiently stiff combination could be found within the project budget, the system will be required to operate between the first and second harmonics during certain tests. To minimize vibration, users will be asked not to operate near the harmonic frequency for an extended period of time. 3.4.6 Mounting Box for Torque Sensor and Two Electric Motors The design of an enclosure was investigated such that it will: protect users from entanglement, allow for proper shaft alignments, and secure the mounting of motors. Several different designs were considered and the selection was narrowed based on: ease of machineability (and thus cost), minimal welding distortion, maximal alignment between motor shafts, and sturdy mounting between motors and floor installation. These are highlighted below in Table 3‐4. 75 omparison forr Different Mou
unting Box Deesigns Table 3‐4: Design Co
Design D
Picture Machineab
bility A Figu
ure Relatiively simple:: cut plate, 3‐1
14 buyy box steel, w
weld up Alignmen
nt Tolerancee Very poorr. Alignment is dependeent on floor,, channelss, boxes, and
d sup
pports Difficult to fit he
Figu
ure eight into Beetter. 3‐1
15 large
e enough mill, little to Alignmen
nt is based on no rroom for socckets for pocket loccation, flipping moun
nting bolts, n
narrow and the piece o
over withoutt a shallow interiorr make it proper jig induces difficcult for torqu
ue sensor alignm
ment errors Figu
ure Simple design allows for Very tigh
ht. Machinistt 3‐1
16 more
e precision. O
Only three must deesign a jig to parts b
but water‐je
et, welding, measurre the piece an
nd CNC macchining from. How
wever the firrst required. Prevventing cut is the alignmentt “pottato chip” warping of pocket, w
which can bee plate
e after welding heat is very precisse and used to very difficcult dimensiion all other geometry from B C Cosst Stock: high Machin
ning: low
w Stock: low ning: Machin
medium Stock: low ning: Machin
mediu
um‐
high Figure 3‐14: Mo
ounting Design
n Concept A. M
Motors Mounteed on Reinforcced Stands with Cantilevered L‐Brackeet Supports 766 d within Square Plate with M
Figure 3‐15: Mountin
ng Design Concept B. Squaree Box Mounted
Motor Mounting Ho
oles oncept C. Squaare Box Rotateed 45 Degrees to Allow for M
Most Figurre 3‐16: Mounting Design Co
Inte
erior Volume ffor Sensor and
d Couplers, Yett Still within M
Motor Mountin
ng Holes Design
n C was the chosen design because it features tthe tightest aalignment to
olerances an
nd allows forr easy accesss to mountinng bolts. Oncce this desiggn was choseen, 777 sttress calculations were p
performed u
using SolidW
Works 3D CAD
D modeling ssoftware and its in
nternal finite
e element an
nalysis solver. Figure 3‐117 shows thee results of tthe simulatio
on. This estimate
ed model sho
owed that lo
oading the stteel mountin
ng plates wo
ould producee a vo
on Mises strress of 3 ksi. This simulattion helped estimate thee thickness o
of the steel walls to
o determine if the motor max torque can damagge the moun
nting box evven under the worse loading
w
g moments. Figure
e 3‐17: Fundam
mental Finite EElement Analyssis of Design C
C with Concenttrated Forces aat the Moto
or Mounting B
Bolt Holes A challenge in Dessign C was allowing for ssufficient tool space for installing th
he motor mount
m
ting bolts. Figure 3‐18 sh
hows the 3D
D CAD modell of Design C
C and clearan
nce fo
or the socket and driver between the box and thhe bolt. Afteer redesignin
ng the box dimensions, tthe proper aamount of sp
pacing for a 99/16” nut an
nd socket waas found. 788 Figure 3‐18: CAD Mo
odel Showing D
Dimensions beetween Motor Mounting Bollt Hole and thee Next Surface
The otther crucial design challenge was heeat stress wh
hen weldingg the plate ends to
o the box. Frrom a design
n perspective
e, welding oon the interio
or of the boxx is ideal. However, we
H
lding technicians can’t gget tools in tthe tight corrners. The manufacturing co
ompany agreed to perfo
orm the weld
ds cautiouslyy, in a staggeered pattern
n to balance the heat. The ressulting deforrmation could then be reemoved durring machining. Particular care waas taken with
h the machinning instructtions. Since the design m
must e motors moounted on opposite end
ds of the boxx, it is produce tightt alignment between the
crritical to assure the align
nment pocke
ets on eitherr end of the box are botth parallel an
nd axxisymmetricc. To minimizze machiningg errors, thee number of machining ssteps prior to milling the ali
m
ignment pocckets was minimized. Thhe box was ccut to length
h (with a sligh
ht amount of sp
pare to accou
unt for the finishing of thhe plates), tthe access ho
ole was cut u
using 799 a water jet, th
he end plate
es were weld
ded on, and then both eends were m
machined to crreate the alignment pocckets. All oth
her features are based o
off of these p
pockets to ensure prope
er alignment. Machining was compleeted incremeentally, utilizing a uniqu
ue jig, built by the m
manufacturin
ng company,, to assure bboth ends off the box were properly aligned. ptions for attaching the mounting box to the flo
oor were Many different op
onsidered ass well. Some
e options inccluded buildiing a framew
work of I‐beaams or box rrails co
to
o slide the te
est rig on. Ho
owever, sincce the rig dooesn’t need tto be interch
hangeable like in most ICE dyn
m
o test stands, direct con
ncrete anchoors could be used. In ord
der to reducee viibrations, ru
ubber isolatio
on dampers were addedd between th
he mountingg box and th
he co
oncrete anchors. Below,, Figure 3‐19
9, is a picturee of the final design featturing the mounting box
m
x, angle braccket, and dampers. Figure 3‐19: 3D CAD Model of Design C wiith Mounting FFeet Brackets and Bushings
800 3.4.7 Powertrain Control Module The powertrain control module (PCM) was supplied by Artisan Vehicle Systems to provide communications between the inverters, batteries, and the user. The PCM is designed to be a plug‐and‐play solution by AVS. However, since this set‐up is not a traditional application where the PCM only receives throttle input from the user, several communication modifications were made. Unlike vehicle control, the inverters in the lab can be controlled in two different methods (speed and torque control). The user then can select a multitude of different options to perform the desired test. These options include holding a set rpm for an extended period of time while incrementally increasing torque. This is very atypical for automotive use but standard for static laboratory testing. 3.4.8 Control Computer The following software was installed on the control computer: LabView, TorqueSense, Parker SSD DSE Lite, and the AVS PCM GUI. It is also equipped with math software to do analysis of results (Excel, MatLab, etc.) as well as modelers for control loops (MatLab Simulink), and 3D CAD modeling for designing layouts and building parts (SolidWorks). Fitted with this software, the computer functions as a way for the user to monitor the system and command the devices to perform in certain desired manners. 3.4.9 Data Acquisition System The data acquisition system (DAQ) chosen for the lab was based on cost, functionality, and expandability. As a main chassis, the National Instruments (NI) 81 CompactRIO (cRIO) with real‐time controller and reconfigurable chassis (NI cRIO‐9074) was chosen because it has: field‐programmable gate array (FPGA), 8 slots for cards, up to 400 MHz real‐time processor, 128 MB DRAM memory, and 256 MB of nonvolatile storage. In addition, the chassis was on sale because it will be replaced soon. One of the disadvantages to this model however, is that it is over‐designed for its purpose in the lab. Since the chassis is FPGA based, this requires additional complex programming, which would have been avoided with a lower performance model. This adds more upfront programming to commission the lab but allows for flexible upgrading of the system in the future. Proper cards were chosen for the cRio to handle the necessary I/O signals (digital in/out and analog in/out), required single power, and sampling rate required for the control system. For digital signals, the NI 9422 (8 channel, 24 to 60 V, 250 μs C series sinking/sourcing digital input module) and NI 9472 (8 channel, 24 V Logic, 100 μs sourcing C series digital output module) modules were chosen. Choice of the analog I/O cards was subject to time delays requirements, which were discussed in detail in section 2.3.4 and 2.4.7. The two analog modules NI 9215 (4 channel, 100 kS/s, 16‐Bit, ±10 V simultaneous sampling C series analog input module) and NI 9263 (4 channel, 100 kS/s, 16‐Bit, ±10 V, analog output module) were chosen because their sampling rate is greater than 1kHz required for AC voltage measurements. Lastly, an NI 9853 2‐Port, High‐Speed CAN Module specifically for NI CompactRIO was selected in order for LabView to have CAN communications with all the key components in the system. This will allow the user to transmit and receive CAN 82 messages between the PCM, BMS, and the inverters. This is necessary for increased flexibility in communications between the devices and greater control of the system. 3.4.10 Emergency Stop Circuit Typically, there are two ways to design an emergency stop (E‐stop) circuit in industrial labs: series and parallel. In the series method, all switches have to be “ok” in order for power to flow to the system. In a parallel design, all switches are independent, but if the user closes one switch then the E‐stop bus is powered causing an emergency stop of the system. The eDTC lab utilizes the parallel system. Once a single emergency stop is signaled (push of an E‐stop switch or digitally pushing the DAQ E‐stop button), 24VDC flows to the trip bus, which immediately powers all shunt trips, enable signals, and contactors to the de‐energized or open state. Figure 3‐20 shows a simplistic flow diagram of the E‐stop circuit design. It is important to note that the E‐stop system is completely independent of the control system and does not require any input from the control computer to operate. However, a signal from the control system can trip the E‐stop system, if necessary. The system is reliable as 24VDC is supplied. Additional redundancy could be built into the system to prevent the system from failing should the 24V become disconnected. However, at this time it was deemed as unnecessary. 83 Figure 3‐20
0: Simplistic Scchematic of E‐sstop Logic 3.5 Check Out and Ve
erification
This fiinal section o
of the dynam
mometer poortion of the eDTC lab discusses the final co
onstruction and implem
mentation of the key com
mponents of the lab (Secction 3.5.1) aas well as an op
w
erational gu
uideline for d
dynamometeer users (Secction 3.5.2). After these two su
ubsections iss the section
n on coursew
work examples (Section 3.6) to show
w how the educational o
objectives off the lab werre met. 844 3.5.1 Imple
ementation aand Constru
uction The complet
T
ted test centter was a successful inteegration of m
motors, inveerters, DC po
ower supply, rege
s
nerative DC bus bar dessign, torque sensor, and liquid coolin
ng circuit. Figure 3‐21 illusstrates the co
onceptual laab layout in aa CAD modeel. The figurees following ( Figure
e 3‐22, Figurre 3‐23, Figure 3‐24, andd Figure 3‐255) illustrate tthe lab as bu
uilt with key com
w
mponents lab
beled. Figure 3‐21: 3D CAD M
Model Detailed
d Layout with C
Component Caall‐Outs 855 Figgure 3‐22: Inve
erter Rack with
h DC Bus Bar C
Cabinet 866 Figure 3‐23: Dyyno Rig with K
Key Componen
nts Called Out Figure 3‐24: Contr ol Cabinet Layyout 877 Figure 3‐25: Ph
F
hysical Layout o
of Constructed
d Componentss 3.5.2 Motor Spin‐Up an
nd Coast Do
own Operatio
on To perform dyno testing, the user must fiirst ensure aall the conneections betw
ween he DC bus baar cabinet, in
nverters, and motors aree properly m
made. All thee E‐stop butttons th
need to be in the open, twisted up, o
off position. The user mu
ust obtain peermission from nteGrid to haave control o
of the AE Summit DC poower supply and power o
on the InteG
Grid In
co
ontrol PC. Using the AVSS PCM GUI software on tthe eDTC co
ontrol compu
uter and DSEE Lite so
oftware, the
e user can co
ontrol the ab
bsorber inve rter and mo
otor. Depend
ding on the sset‐
up there can be two ways to operate
e the dyno. O
One way is to
o have an ad
dditional con
ntrol omputer to communicatte solely witth the unit u nder test motor and invverter via thee co
DSE Lite softw
D
ware and maanual contro
ol. This methhod could bee substituted
d with a virtu
ual 888 PC application that allows two DSE programs to run simultaneously on one PC. The other method is to have the NI cRIO DAQ communicate via CAN to the PCM and inverters, which can, in turn, communicate with both inverters. The first method is the only implemented method at this time. To start the system, the user must set the InteGrid DC power supply control computer to command a bus voltage of 350 VDC. At the eDTC control computer, the user may turn the enable key to begin testing the dyno. While monitoring the program on the eDTC control computer (either LabView, AVS PCM GUI, or DSE Lite) the user can define speed set points on the absorber and then “throttle” up or down the torque request of the UUT motor. While performing different tests, the user can opt to record data via the PCM GUI, which will save data as a .csv file for a later use. When the user is finished performing the desired tests they must spin the motors down to a stop by reducing the set points and switching the enable switch to the OFF position for both inverters. After this, the InteGrid control PC must properly shut down the DC power supply and open the service disconnects for all power equipment. 3.6 Course Work Examples This section illustrates four different labs that professors will use to educate students using the eDTC laboratory. These labs include mapping the torque of the motors (Section 3.6.1), finding the efficiency of the lab (Section 3.6.2), understanding thermal effects (Section 3.6.3), and an introduction to grid‐attached‐storage (Section 3.6.4). 89 3.6.1 Lab #1: Map Torque versus Speed Curve Introduction: The objective of this lab is to learn more about motor performance, in particular, the torque curve. Students will experience creating a torque map similar to the torque curve in Figure 3‐26. To perform this test, the user defines a set speed for the absorber motor and identifies the maximum torque achievable by UUT at that speed. Since the motors are directly coupled, both will operate at the same speed at all times, except for minor vibrational variations. The UUT is operated under torque control – i.e. the user commands the UUT to output a defined torque. The absorber is operated under speed control – i.e. it is commanded to always maintain a set speed. When the user changes the UUT’s commanded torque, the UUT attempts to produce that torque, while the absorber attempts to maintain a set speed. Since the absorber is controlled so that it can always produce more torque than the UUT, the speed may vary briefly before quickly returning to the commanded level. During this test, the user will define 16 different speeds and record the max torque that results for each. Students must plot these max points and compare them to the torque curve below courtesy of www.artisanvs.com. 90 Figure 3‐26: Torrque Curve at CContinuous an
nd Peak Torque Learning Objectivves: 1. Understan
nd dynamom
meter controols 2. Learn how
w to measure
e the outputt torque of aa motor 3. Gain experience with an electric m
motor torquee curve and performancce Lab O
Operations: 1. Checkk with the lab
b instructor that all the proper conn
nections are made betw
ween motorrs, inverters,, and DC buss cabinet. Pluug‐in the torrque sensor power, Eagle 12V CPU power, aand the LPS power plugss. Ensure all cabinets and
d covers aree closed
d, E‐stop swiitches in the
e OFF positioon (up), pum
mps and fans are turned on and w
water is flowiing, and nothing is obstrructing the sspinning mottion of the motorr shafts 2. In the
e InteGrid PV
V Simulator aarea, switch on the power and log in
nto the control computer. Open P
PV_Manual_
_Operator.vii (found on tthe desktop)) and follow the 911 powering on instructions (screenshot shown in Figure 3‐27). Set the command voltage to 350VDC 3. Return to the eDTC. Open the Motor_Map.vi (example screen shot is shown in Figure 3‐28), SSD Drives: DSE Lite, and Calmotors GUI (also known as AVS PCM GUI) 4. Set the absorber speed to 1,000rpm and switch on the enable switch on the desk control box 5. Define the UUT inverter torque to 10% and switch on the enable switch 6. Quickly accelerate to 200% torque and let the motor stabilize for five seconds and record the torque value 7. Ramp the torque back down to 0% 8. Change the absorber speed to 1,500rpm 9. Monitor the winding temperature at all times to assure that it is within 5oC of the initial winding temperature. If the temperature increases, wait until it cools sufficiently 10. Quickly accelerate the UUT to 200% torque and let it stabilize for 5 seconds. Record the torque at this point 11. Ramp the torque back down to 0% 12. Repeat steps 6‐9 with 500rpm increments until you reach 8,000rpm 13. Plot your results on a Torque vs. Speed graph and submit to the instructor 92 Figgure 3‐27: PV SSimulator Screeenshot of Manual_Voltage.vi Figgure 3‐28: Scre
eenshot and Exxplanations off Motor_Map.vvi Home
ework: Now that you have seen how
w a dyno works, explain what is goin
ng on in
n “real world
d” applicatio
ons. Explain tthe analogy between the absorber aand road loaad. During the te
D
esting of the motor, whyy did the torqque drop offf at a certain
n speed? Exp
plain 933 why electric motors behave in this manner. Are there different ways to control electric motors to change the impacts of the motor torque curve? Explain your reasoning. Hand in your plot from in‐lab and your essay answers to the questions above to your lab instructor before the beginning of the next lab. 3.6.2 Lab #2: Map Motor Efficiency Lab Introduction: The purpose of this lab is to better understand the efficiencies of an electric motor. As you learned in‐class, there are many parts of an electric motor, which are subject to friction, heat, and electrical losses. This lab exercise investigates these loses at a system level. To investigate efficiency, you will compare the electrical power input into the UUT inverter, the mechanical power transmitted to the absorber via the shaft coupling, and the electrical power returned to the system by the absorber. Given measurements at three points, it is possible to isolate the efficiency of subsets of the laboratory: The efficiency of UUT inverter/motor, and the efficiency of the absorber inverter/motor. Since one motor is operating exclusively in traction mode, and the other exclusively in regenerative mode, it is possible to assess the efficiency of the same motor model in both modes simultaneously. Figure 3‐29 shows the estimated efficiencies for each component in the system. 94 Figure
e 3‐29: Flow Schematic with
h Efficiencies at Each Component Your ggoal of this lab is to meaasure actual losses, explaain their origgin, map how
w effficiency varries with spe
eed and torq
que, and whaat impact effficiency has on vehicle operation. To
o complete tthis, you will map the effficiency of th
he absorberr or UUT. Figure 3‐30 (courtessy of http://w
www.uqm.com/propuls ion_specs.p
php) is an exaample moto
or will produce at the end oof the lab. effficiency maap that you w
955 Figure 3‐‐30: Example M
Motor Efficien
ncy Map Learning Objectivves: 1. Understan
nd the power flow withinn the lab com
mponents 2. Get a hand
ds‐on perspe
ective of effiiciencies of aa motor 3. Make a mo
otor efficien
ncy map In most test labs tthe power fo
or the dyno motor is typ
pically suppliied from a battery pack to simulate actual vehiccle testing. H
However, in tthe case of tthe eDTC lab
b, the power comess from a controllable DC power suppply. The DC ssupply sets tthe DC bus vo
oltage seen by both the UUT and ab
bsorber inve rters, in much the samee way that a battery pack sets the DC bus voltage in a vehicle.. Since the aabsorber is ““regeneratin
ng” up” power ffor losses in tthe power onto the DC bus, tthe DC supply only proviides “make u
966 UUT and absorber. By regenerating the absorbed power back onto the DC bus, the lab minimizes the energy consumed and reduces the size of certain components. Lab Procedure: 1. Check with the lab instructor that all the proper connections are made between motors, inverters, and DC bus cabinet. Plug‐in the torque sensor power, Eagle 12V CPU power, and the LPS power plugs. Ensure all cabinets and covers are closed, E‐stop switches in the OFF position (up), pumps are turned on and water is flowing, and nothing is obstructing the spinning motion of the motor shafts 2. In the InteGrid PV Simulator area, switch on the power and log into the control computer. Open PV_Manual_Operator.vi (found on the desktop) and follow the powering on instructions. Set voltage to 350VDC – the emulated battery voltage – and the safety current limit to 50A 3. Return to the eDTC control computer. Open the Motor_Map.vi, SSD Drives: DSE Lite, and Calmotors GUI (also known as AVS PCM GUI) 4. Set absorber speed to 1,000rpm and then switch the enable to on for the absorber inverter 5. Record winding temperature 6. Define the UUT inverter torque to 10% and switch the enable switch to the on position 7. Enable data logging 97 8. Slowly ramp up the UUT torque to 200% using intervals of 20% and pausing for 5 seconds between steps (the more systematic you are with this, the easier it will be to interpret your data) 9. Ramp torque down to 0% 10. Check the winding temperature. Is it within 5oC of the starting temp? If not, wait until it cools (the cooling loop should quickly cool the motors) If yes, proceed to the next step 11. Change the absorber speed to 1,500rpm 12. Repeat step 5 13. Ramp down torque to 0% 14. Repeat steps 5‐7 until absorber speed is 8,000rpm 15. Ramp down torque to 0% and absorber speed to 0rpm 16. Switch off enable switches for both inverters, safely power down the DC power supply according to instructions, switch off all break disconnects 17. Open the data file showing absorber speed at each: torque, speed, voltage and current (for both inverters) of the system Homework: Do the calculations using equation one (show below) to determine the system efficiency at each data point and create a plot of Torque vs. Speed with “islands” of efficiency in different colors like the example in Figure 3‐30. 98 3.6.3 Lab #3
3: Investigatte Temperatture Effects on Electric M
Motors Introd
duction: The
e purpose of this lab is too investigatee the impactt of the wind
ding te
emperature on the perfo
ormance of a motor opeerating in diffferent statees. You will determine ho
ow changes in winding re
esistance im
mpact the mo
otor output in four sccenarios: ste
eady state co
old, steady sstate hot, ac celeration, aand regenerration. Be prepared to aanswer wherre the heat iis being trannsferred to aand why. Learning Objectivves: 1. Understan
nd the impacct of temperrature on mo
otor perform
mance 2. Get a hand
ds‐on perspe
ective of mootor operatio
on in different condition
ns Lab Prrocedure: 1. Checkk with the lab
b instructor that all the proper conn
nections are made betw
ween motorrs, inverters,, and DC buss cabinet. Pluug‐in the torrque sensor power, Eagle 12V CPU power, aand the LPS power plugss. Ensure all cabinets and
d covers aree closed
d, E‐stop swiitches in the
e OFF positioon (up), pum
mps are turneed on and w
water is flow
wing, and no
othing is obsttructing the spinning mo
otion of the motor shaftts 2. In the
e InteGrid PV
V Simulator aarea, turn onn the power and log into
o the control computer. Open P
PV_Manual_
_Operator.vii (found on tthe desktop)) and follow the powering on instrructions. Sett voltage to 3350VDC 3. Open the Motor_
_Map.vi, SSD
D Drives: DSEE Lite, and Caalmotors GU
UI (also know
wn as AVS PCM GUI) 999 4. Set the absorber speed to 1,000rpm, switch the enable absorber switch to on, and record the initial winding temperature 5. Switch the UUT inverter enable switch to on after setting the torque to 5% 6. Slowly accelerate to 75% torque and wait until the system stabilizes 7. Note the temperature over 1 minute 8. Bring the torque down to 0% 9. Quickly accelerate to 75% torque and wait until the system stabilizes and record the temperature in both the UUT and absorber motors 10. Bring the torque back down to 0% 11. Let the windings cool down to the initial temp before moving on to the next step 12. Change the absorber speed to 5,000rpm 13. Quickly accelerate to 200% torque 14. Record how the temperature of the winding responds to this input over time 15. After about 30 seconds, return to the torque to 0% and the absorber speed to 1,000rpm 16. Note the change you see in the temperature 17. Return both torque and speed to 0 and wait for the system to come to a stop. Let the coolant flow until the winding temperature is at the initial temperature when you started the lab Homework: Discuss the effects of time and load in terms of temperature in the UUT and absorber motors. Present the data you collected in a manner that you best think describes the heat transfer. 100 Consider automakers performing these tests to determine how to best cool the motor. When do you expect the motors to experience different heating characteristics? Low torque up a hill, driving in reverse, highway cruising? Describe what will happen in each one of these cases and a few of your own based on the results you have gathered from this lab. Can you model the temperature characteristics of the motor at different speeds and at different length durations? Organize these findings and hand them into the lab director before the start of the next lab. 3.6.4 Lab #4: Intro to Grid‐Attached‐Storage Lab Introduction: The purpose of this lab is to give students an initial contact with grid‐attached‐storage to understand the relationship between grid demand and allowable battery charge and discharge rates. Students will experience how a storage device, such an automotive pack battery, interfaces with the grid. “Grid‐attached‐storage” (GAS) involves connecting a storage device – in this case a battery – to the utility grid and controlling the charge and discharge rate of the storage to meet objectives such as grid stabilization, frequency regulation or peak reduction. For example, a battery could be utilized for peak shaving – charging during periods of low electricity consumption (typically when electricity cost is low) and discharging during periods of peak electricity consumption (typically when electricity cost is high). 101 This lab focuses on the basic requirements to control a battery GAS system connected to the power grid. Students will monitor the behavior of the battery while changing commanded charge and discharge levels. Learning Objectives: 1. Understand grid‐attached‐storage components 2. Get a hands‐on perspective of GAS current control methods 3. Witness the effects of battery performance with short pulse demands Lab Procedure: 1. Open the PCM GUI and DSE Lite programs at the eDTC control PC 2. Check with the lab supervisor that all the proper connections between the inverter, battery, and transformer have been made. Ensure all safety checks are in place: no cables are improperly color matched, everything is properly connected, all shields and covers are closed, all E‐stop switches are in the off position 3. Plug in the GAS water pump and close the priming valve to properly activate the cooling loop. Plug in the Eagle 12V power supply plug and turn on the GAS electric fan at the desk control box 4. Check that the battery is properly communicating with the PCM via CAN 5. Before moving to the next step, ensure that the LC filter pre‐charge e‐stop switch in the LC filter cabinet is in the off or twisted up position. Strict adherence to this will reduce the risk of arcing and damage to the LC filter 6. Follow InteGrid connection instructions to energize the 480VAC main tie breakers 102 7. Close the 480VAC I‐Beam breaker on, turn on the LC filter fans, and close the LC filter knife switch on, letting the pre‐charge circuit come online 8. Turn on the inverter through the ENABLE rocker switch via the control box by the eDTC control PC 9. Turn the battery key‐on at the control desk box to the on position. You should hear the battery contactors switch the battery pre‐charge on and then close the rest of the contactors 10. Once the system is synchronized with the grid, manual current control can begin 11. Using DSE Lite increase or decrease the current set‐points in 1% increments following the control provided by your instructor 12. While manually controlling the current set points, monitor the battery SOC to assure that the battery pack remains within its operational SOC limits 13. Take notes of how the controller responds, battery pack behaves, and the time required for the system to respond to commanded changes in current output 14. Power down the system by commanding the PCM to open the battery contacts, turning the key to the off position, disabling AFE (absorber inverter ENANBLE rocker switch off), opening/switching off the LC filter breaker, and lastly the 480VAC breaker 15. To safely store the lab: unplug the battery connections, communication cable (Deutsche 23‐pin connector). Unplug all the other power devices in the eDTC lab (water pumps, Eagle 12VDC CPU power supply, linear power supply (LPS), etc.) 103 Homework: Hand in a discussion and a controller model to your lab supervisor before the start of the next lab. In your discussion, consider the following questions: 
How do the asymmetric charge and discharge limits impact the control algorithm for the GAS system? 
Given that a fully charged or discharged battery system can no longer provide full control functionality to the grid, what control architecture would you propose to maintain the SOC of the battery while still responding to grid control requirements? 
Given the step response characteristics of the system, what is the maximum allowable control frequency for the system? One application of GAS is to “firm” output from a renewable power plant, such as a wind farm or photovoltaic system. The concept behind firming is to control charging/discharging of the GAS system so that the combined output of the renewable plant and GAS system maintains a constant, forecasted value. The second part of this homework is to design a controller for renewable firming. As you witnessed, the lab currently uses manual set‐point control for real and reactive power commands. Your objective is to design a control algorithm to firm the output from a renewable power plant. Your instructor will provide a time series output from a renewable power plant. In the next lab, you will test your algorithm simulation and controller (pending instructor approval) by uploading it to the inverter software controller to verify its robustness. 104 4.
Summary and Conclusions The eDTC lab at Colorado State University is a test bed that advances the knowledge and training of engineers for a sustainable future. It allows for educating the next work force in electrical mobility and provides researchers with the components to test the implications of grid‐attached‐storage. The results from the lab will inform how utilities and electrified vehicles interact as well as the repurposing EV batteries for GAS systems. This thesis described the need for understanding grid storage and its implications, what challenges are accompanied with designing a test bed of this scale with automotive‐grade components, why different components were chosen over alternatives, and a brief demonstration of the system operation. The goals of the design were accomplished and a successful lab was constructed. 4.1 Future of the eDTC This is an exciting time for the eDTC lab because there are several new uses designated for the GAS and dyno portions of the lab. First, new graduate students will design new control algorithms for GAS for testing in both simulation and class environments. Second, used batteries from Altairnano and A123 were donated to CSU specifically for battery second use testing at the eDTC. Testing protocols and performance of these batteries is of great importance to industry and national labs. 105 Third, EcoCar2 teams will begin drive cycle testing using the dyno portion of the eDTC. This will provide real world results and validation to the modeling currently being evaluated. Finally, beginning spring 2011, the new ENGR580a1 class will utilize the course exercises designed in this thesis. With these and many more exciting projects, the eDTC has begun to show its importance and usefulness to both researchers and university students. 106 WORKS CITED Blooming, T., & Carnovale, D. (2007). Application of IEEE Std 519‐992 Harmonic Limits. IAS Atlanta Section. Appleton, WI: IEEE. Bredenberg, A. (2011, April 25). Battery Storage: Smoothing Out the Wrinkles of Renewable Energy. Retrieved May 22, 2011, from Thomas Net News: http://news.thomasnet.com/green_clean/2011/04/25/battery‐storage‐smoothing‐out‐
the‐wrinkles‐of‐renewable‐energy/ Brockman, J. (2008, July 22). Plug In, Drive On: The Promise of Next‐Gen Hybrids. Retrieved May 14, 2011, from NPR: http://www.npr.org/templates/story/story.php?storyId=92792068 Brooks, A. (2002). Vehicle‐to‐Grid Demonstration Project: Grid Regulation Ancillary Service with a Battery Electric Vehicle. AC Propulsion. CARB. Bureau of Transportation: Statistics. (2010, May 03). Number of U.S. Aircraft, Vehicles, Vessels, and Other Conveyances. Retrieved May 12, 2011, from Research and Innovative Technology Administration (RITA), BTS: http://www.bts.gov/publications/national_transportation_statistics/html/table_01_11.
html Coomb, H. S. (2011). Renewable Sustainable Expeditionary Power. U.S. Navy, Office of Naval Research. Arlington, VA: U.S. Navy. Cready, E., Lippert, J., & Phil, J. (2003). Technical and Economic Feasibility of Applying Used EV Batteries in Stationary Applications. Sandia National Laboratories. Albuquerque, NM: Department of Energy. Demand Energy. (2011). Demand Energy ‐ Renewables Firming. Retrieved September 2011, from Demand Energy: http://demandenergynetworks.com/applications/use‐case‐renewables‐firming.asp Denholm, P., & Short, W. (2006). An Evaluation of Utility System Impacts and Benefits of Optimally Dispatched Plug‐In Hybrid Electric Vehicles. National Renewable Energy Laboratory. Golden, CO: Department of Energy. 107 Dickinson, R., & Milano, S. (2002, July 11). "Isolated Open Loop Current Sensing Using Hall Effect Technology in an Optimized Magnetic Circuit. Concord, NH, USA. Duvall, M. (2008). How many plug in hybrids can a smart grid handle? Plug‐In 2008. San Jose, CA: EPRI. EPRI. (2011). Plugging In: A Consumer's Guide to the Electric Vehicle. Electric Powertrain Research Institute. Palo Alto, CA: EPRI. Eyer, J., & Corey, G. (2010). Energy Storage for the Electricity Grid: Benefits and Market Potential Assessment Guide. Sandia National Laboratory. Albuquerque, NM: Department of Energy. Gavine, A. (2011, January). Ticket to Ride: User‐Friendly System. Electric and Hybrid Vehicle Technology International , p. 50. Goggin, M. (2008). Wind Power and Energy Storage Fact Sheet. American Wind Energy Association. AWEA. Gupta, P. (2009, August 26). Utility wants to deploy largest grid battery ever. Retrieved May 30, 2011, from Reuters: http://www.reuters.com/article/2009/08/26/us‐
edison‐battery‐idUSTRE57P4PJ20090826?sp=true Hadley, S., & Tsvetkova, A. (2008). Potential Impacts of Plug‐in Hybrid Electric Vehicles on Regional Power Generation. Oak Ridge National Laboratory. Oak Ridge, TN: Department of Energy. Himelic, J., & Novachek, F. (2010). Sodium Sulfur Battery Energy Storage and Its Potential To Enable Further Integration of Wind,. Xcel Energy Renewable Development Fund. Denver, CO: Xcel Energy. Hoevenaars, T. (2003). Interpreting IEEE Std 519 and Meeting its Harmonic Limits in VFD Applications. IEEE. Jost, B. (2011, September 1). Who Pays for Firming Up Variable Energy Resources? Retrieved September 2011, from http://www.powermag.com/renewables/wind/Who‐Pays‐for‐Firming‐Up‐Variable‐
Energy‐Resources_3977.html Kempton, W., & Tomić, J. (2005). Vehicle‐to‐grid power fundamentals: Calculating capacity and net revenue. Journal of Power Sources , 144 (1), 268‐279. 108 Kempton, W., Tomić, J., Letendre, S., Brooks, A., & Lipman, T. (2001). Vehicle to grid power: Battery, hybrid and fuel cell vehicles as resources for distributed electric power in California,. University of California, Davis, Davis Institute for Transportation Studies. Davis, CA: University of California. Kempton, W., Udo, V., & Huber, K. (2008). A Test of Vehicle‐To‐Grid (V2G) for Energy Storage and Frequency Regulation in the PJM System. Mid‐Atlantic Grid Interactive Cars Consortium. Kintner‐Meyer, M., Schneider, K., & Pratt, R. (2006). Impacts assessment of plug‐
in hybrid vehicles on electric utilities and regional US power grids, Part 1: Technical Analysis. Pacific Northwest National Laboratory. Richmond, WA: PNNL. Kirby, B. (2004). Frequency Regulation Basics and Trends. Oak Ridge National Laboratory. Oak Ridge, TN: Department of Energy. Levitan, D. (2010, October 14). Rising Hopes that Electric Cars Can Play a Key Role on the Grid. Retrieved May 22, 2011, from Yale Environment 360: http://e360.yale.edu/feature/rising_hopes_that_electric_cars_can_play_a_key_role_on
_the_grid/2328/ Lutz, M., Zimmerle, D., Huff, B., & Bradley, T. (2011). Design and Construction of Grid Attached Storage Simulator. ASME 2011 5th International Conference on Energy Sustainability & 9th Fuel Cell Science, Engineering and Technology Conference (pp. 1‐9). Washington, DC: ASME. McGuinness, M., & Fields, C. (2011, May 18). Majority of Consumers Ready to Consider Buying Plug‐in Electric Vehicles, But Challenge Utilities with their Car Charging Demands. Retrieved May 21, 2011, from Accenture: http://newsroom.accenture.com/article_display.cfm?article_id=5205 Miller, N., Zrebiec, R., Delmerico, R., & Hunt, G. (1996). Battery energy storage systems for electric utility, industrial and commercial applications. Battery Conference on Applications and Advances, 1996., Eleventh Annual (p. 235). Long Beach, CA: Dept. of Power Syst. Eng., General Electric Co. Mitlitski, J. (2011, February 18). Raising the Volt‐Age: Is Obama's Goal of 1 Million Electric Vehicles on U.S. Highways by 2015 Realistic? Retrieved May 12, 2011, from Scientific American: http://www.scientificamerican.com/article.cfm?id=one‐
million‐electric‐vehicles‐by‐2015 109 Moura, F. (2006). Driving Energy System Transformation With “Vehicle‐to‐Grid” Power. International Institute for Applied Systems Analysis. Naughton, K. (2011, January 10). Ford to Hire 7,000 Workers to Help Manufacture Electric, Hybrid Vehicles. Retrieved May 12, 2011, from Bloomberg: http://www.bloomberg.com/news/2011‐01‐10/ford‐to‐hire‐7‐000‐workers‐to‐help‐
manufacture‐electric‐hybrid‐vehicles.html Neubauer, J., & Pesaran, A. (2010). HEV/EV Li‐Ion Battery Second‐Use Project. National Renewable Energy Laboratory. Golder, CO: Department of Energy. Neubauer, J., & Pesaran, A. (2010). Secondary Use of PHEV and EV Batteries – Opportunities & Challenges. 10th Advanced Automotive Battery Conference (pp. 1‐14). Orlando, FL: Department of Energy. Norton, R. (2000). Machine Design (2nd ed.). Prentice Hall. Pacific Gas and Electric. (2011). Pacific Gas and Electric ‐ Tarrifs. Retrieved September 2011, from http://www.pge.com/tariffs/tm2/pdf/ELEC_SCHEDS_A‐10.pdf Parks, K., Denholm, P., & Markel, T. (2007). Costs and Emissions Associated with Plug‐In Hybrid Electric Vehicle Charging in the Xcel Energy Colorado Service Territory. National Renewable Energy Labortory Report. Golden, CO: Department of Energy. Quinn, C. (2011). Evaluation of Distributed Energy Storage for Ancillary Service Provision. MS Thesis, Colorado State University, Mechanical Engineering, Fort Collins, CO. Quinn, C., Zimmerle, D., & Bradley, T. (2009). The Effect of Communication Architecture on the Availability, Reliability and Economics of Plug‐in Hybrid Electric Vehicle‐to‐Grid Ancillary Services. Journal of Power Sources , 195, 1500–1509. Reiter, C. (2010, 12 June). BMW Seeks 2,600 New Employees by End of 2011 on Electric, Hybrid Systems. Retrieved May 12, 2011, from Bloomberg: http://www.bloomberg.com/news/2010‐12‐06/bmw‐seeks‐2‐600‐new‐employees‐by‐
end‐of‐2011‐on‐electric‐hybrid‐systems.html Sandia National Laboratories. (2010). Electric Power Industry Needs for Grid‐
Scale Storage Applications. Sandia National Laboratories. Department of Energy. Sensor Technology. (2011). TorqSense ‐ How it Works. (Sensor Technology) Retrieved June 11, 2010, from Sensor Technology: http://www.sensors.co.uk/about‐
us/technology/torqsense/ 110 Sheets, A. (2011). The ReGenerator: Alternative Energy for Expeditionary Missions. Retrieved May 23, 2011, from Raytheon: Technology Today: http://www.raytheon.com/technology_today/2011_i1/regenerator.html Short, W., & Denholm, P. (2006). A preliminary assessment of plug‐in hybrid electric vehicles on wind energy markets. National Renewable Energy Laboratory. Golden, CO: Department of Energy. Sioshansi, R., Denholm, P., Jenkin, T., & Weiss, J. (2009). Estimating the value of electricity storage in PJM: Arbitrage and some welfare effects. Elsevier , Energy Economics, 269–277. Slocombe, J., Wagner, L., Heber, A., & Harner, J. (1990). Test Equipment for Determining Performance of Electric Motors. Applied Engineering in Agriculture , 100. Tomić, J., & Kempton, W. (2007). Using fleets of electric‐drive vehicles for grid support. Using fleets of electric‐drive vehicles for grid support , 168 (2), 459‐468. Turton, H., & Moura, F. (2008). Vehicle‐to‐grid systems for sustainable development: An integrated energy analysis. 75 (8), 1091‐1108. UC Davis. (2010). Second Life Applications and Value of “Traction” Lithium Batteries. University of California at Davis, Plug‐In Hybrid Electric Vehicle Research Center. Davis, CA: UC Davis. Uhlenbrock, R. e. (2010, May). Electric Drive Train Test – Quo Vadis? ATZ Autotechnology , 10, pp. 56‐9. Vazquez, S., & Lukic, S. e. (2010, December). Energy Storage Systems for Transport and Grid Applications. IEEE Transactions on Industrial Electronics , 57 (12), pp. 3881‐3895. Walawalkar, R., & Apt, J. (2008). Market Analysis of Emerging Electric Energy Storage Systems. National Energy Technology Laboratory. Pittsburgh, PA: Department of Energy. Warrington, R., Kirkpatrick, A., & Danielson, S. (2010). Vision 2030 ‐ Creating the Future of Mechanical Engineering Education Phase 1 Report. ASME Center for Education Task Force. Williams, B. (2010). A Strategy for Overcoming Plug‐in‐hybrid Battery Cost Hurdles in California: Integrating Post‐Vehicle Battery‐to‐Grid (B2G) Use Values. TRB 111 Annual Meeting. 538, pp. 1‐28. Washington, DC: University of Califronia: Berkley: Transportation Sustainability Research Center. Wolkin, M. (2011, February 1). Before Recycling, Electric Car Batteries Enjoy a Second Life. Retrieved February 11, 2011, from Conversations for Accelerating the Transition to Sustainable Transportation: A Better Place Blog: http://blog.betterplace.com/2011/02/before‐recycling‐electric‐car‐batteries‐enjoy‐a‐
second‐life/ Zhou, D. (2003, August 01). Testing Open and Closed Loop Current Sensors. Los Gatos, CA, USA. 112 Appendix A– Wiring Schematics
113 114
Figure 0‐1: Control Cabinet W
Wiring 115
Figure
e 0‐2: DAQ Wiring
116
Figure 0‐3: Lin
near Power Supplyy Wiring 117
Figure 0‐4: Powertrain Control Mod
dule Wiring 118
Figgure 0‐5: E‐Stop, B
Breakers, and Lightt Tower Wiring 119
Figure 0‐6: GFC
CI and Battery Pacck Wiring 120
Figure 0‐7: Contro
ol Desk and Computer Wiring 121
Figure 0‐8: GAS Set‐Up Wirring 122
Figure 0‐9
9: Dyno Set‐Up Wiring 123
Figure 0‐10: Eagle 5
500W CPU Power SSupply Wiring 124
Figure 0‐1
11: Electric Fans Wiring
125
Figure 0‐12: Terminal Block Layout in Controll Cabinet 126
Figure 0‐‐13: Terminal Blockk Layout in Control Cabinet (continued) 127
Figurre 0‐14: Wiring Bun
ndle Distribution in
n Control Cabinet
Appendix B– Artisan Vehicle Systems Pin‐Out Listings 128 Table 0‐1
1: PCM 35 Pin Assignments
129 Table 0‐2
2: PCM 25 Pin Assignments
Table
e 0‐3: Absorberr Inverter Feed
dback Cable Asssignments forr Dyno and GA
AS Set‐Ups 130 Table 0‐4: UU
UT Feedback Cable Pin Assignments
131
Table 0‐5: Deu
T
tsche Pin Batttery Connectorr Assignments Taable 0‐6: Colorred Bundle Asssignments and
d CAN Labelingg 1332 Appendiix C– GD&TT Print for Motor Mo
ounting Bo
ox Shown below is th
he geometric dimensionns and toleraances for thee motor mounting box
m
x. 1333 Appendix D–Natural Frequency MatLab Procedure In order to further narrow down the results, an investigation of the natural frequency of the sensor and couplings in the system must be analyzed. The motor rotates at a maximum speed of 8,000rpm and thus the natural frequency needs to be above this in order to not experience resonance. According to Norton, “usual design strategy is to keep all forcing, self‐exciting, frequencies below the first critical frequency by a factor of three or four” (Norton, 2000). Therefore it is crucial to determine the system natural frequency of the torque sensor and couplings to be greater than 32,000 rpm for safe operation. If the shaft system cannot be made higher than the required rotating frequency, then the motors must be able to accelerate rapidly enough through the resonance before the vibrations have a chance to build up amplitude. This allows the system to operate at a speed higher than resonance. Figure 0‐15 (Source: www.me.mtu.edu) illustrates the excitation at different dampening values for different frequency ratios. The purpose of this plot is to show that as the natural frequency of the system is approached (/n = 1), the displacement amplitude of the oscillations increases significantly depending on the zeta (dampening) value. 134 Figure 0‐15: Naatural Frequen
ncy Amplitude Response Plott Utilizing Matlab, tthe differentt natural fre quencies of the system for each torrque se
ensor and co
oupling can b
be determin
ned. The diffferent mass moment of inertias and to
orsional spring rates were inputted into an m‐fille, which deetermined th
he natural frrequency of the system with each se
ensor. The m
m‐file represented the syystem as a sp
pring‐mass ssystem show
wn below in Figure 0‐16. Figure 0‐16: SSpring‐mass diagram for thee Sensor Tech aand R+W coup
plings set‐up 1335 Each mass moment of inertia (J) as well as torsional stiffness (K) is represented for the motor and motor shaft, coupling, torque sensor, and identical coupling and motor. The caveat to this analysis is how to model the spring representation of the coupling torsional stiffness. It can be modeled on either the motor shaft side or torque sensor shaft side of the coupler. Therefore an investigation will first be made between these two different scenarios defined as caveat #1 and caveat #2. Figure 0‐16 uses arrows to highlight how the coupling spring could be modeled on either side. These two caveats are to be taking into account when analyzing the final natural frequency results. However, it is crucial to understand that there needs to be a spring between two masses so there must always be at least one coupler spring on one side of the torque sensor in order for the analysis to run. With this in mind, only the left side coupler spring was moved between caveat #1 and caveat #2 comparisons, leaving the right side coupler spring connected to the torque sensor side. Using the tabulated values from Table 3‐2 and Table 3‐3, the first investigation (Sensor Tech RWT320 and R+W BK2 couplers) can be assigned to the above variables in Figure 0‐16 for the scenario of the stiffest coupling and cheapest torque sensor (R+W BK2 and RWT320 respectively). Below are the MatLab variable constant definitions to the variables seen in Figure 0‐16. %Mass Moments of Inertia:
Jm = 6.271E-3; %kg*m^2 (motor rotor) Artisan GP150WC
Js = 1.295E-4 %kg*m^2 (motor shaft) calculated using
stainless steel and actual shaft mass
Jc = 4.5E-3; %kg*m^2 (coupler) assuming R+W America
BK2 coupler
136 Jts = 1.431e
e-4; %kg*
*m^2 (tor
rque sens
sor) assu
uming
RWT320 Torque sensor
s
%Torsiona
al Stiffn
ness (spr
ring cons
stants):
Ks = 1.33E5; %Nm/rad
d (motor shaft) c
calculate
ed assumi
ing
stainl
less stee
el
Kc = 1.91E5; %Nm/rad
d (BK2 co
oupler) g
given
Kt = 3.56E4; %Nm/rad
d (torque
e sensor) given
The Kss (motor shaaft torsional stiffness) w
was calculated using G (M
Modulus of Rigidity) = 73X109 Pa for sstainless steel and the toorsional stifffness equation: where
e der to calculaate the shaftt mass, the eequation: In ord
∗ , w
where is deensity and is volum
me. The den
nsity of stainless steel waas assumed to be 7.90X106 g/m3 forr 304 sttainless stee
el and the vo
olume was caalculated us ing Using these value
es, the Ks forr stainless steeel of the m
motor shaft can be calculated ass m‐file relatess the variablees to the sprring Once all the variables are assiigned, the m
mass system m
shown in Figgure 0‐16. N
Note that som
me mass meeasurementss appear to b
be in 1337 kg, while others are in grams. These were converted to a common base in the calculation section of the simulation. % Relate the physical values to the generic spring mass
system
J1 = Jm+Js
J2 = Jc;
J3 = Jts;
J4 = Jc;
J5 = Jm+Js;
% Caveat #1: Assuming the coupler "spring" is on the
torque sensor shaft "spring"
K1 = Ks;
K2 = 1/(1/Kc + 1/Kt);
K3 = Kc;
K4 = Ks;
However if both the caveats mentioned above are considered, the program can be re‐run with the coupler spring on the other side with the code shown below. % Caveat #2: Assuming the coupler "spring" is on the
motor shaft "spring"
K1 = 1/(1/Kc + 1/Ks);
K2 = Kt;
K3 = Kc;
K4 = Ks;
138 Continuing with the program code, the definition of the matrices for the Eigen vector of phi and lambda appear as:
% Define the matrices of coefficients
J_mat = [J1 0
0
0
0;
0
J2 0
0
0;
0
0
J3 0
0;
0
0
0
J4 0;
0
0
0
0
K_mat = [ K1 -K1
-K1
J5]
0
K1+K2 -K2
0
-K2
0
0
-K3
0
0
0
0
0;
0
0;
K2+K3 -K3
K3+K4
-K4
0;
-K4;
K4]
[phi,lambda]= eig(K_mat,J_mat)
for i=1:5
nat_freq(i) = sqrt(lambda(i,i));
end
nat_freq_rpm = nat_freq*60/(2*pi)
139 The re
esults from tthese calculaations are taabulated in TTable 0‐7. Reemember that Caveat #1 is w
when the torque sensor and couplinng springs arre together. Caveat #2 iss when the cou
w
upler spring is paired witth the motor shaft sprin
ng (on the lefft side of thee to
oque sensor). Table 0‐‐7: Sensor Tech
h with R+W Co
ouplers Caveatt Comparison Results Natural Frequencyy At: Mo
ode 2 Mo
ode 3 Mo
ode 4 Mo
ode 5 Caveatt #1 n (rpm)
19,5
500 67,5
570 72,4
450 379,,920 Caveat #2 n (rpm)
20,1160 55,1130 70,9950 384,5510 Differeence (#1‐##2) ‐660
0 12,44
40 1,50
00 ‐4,59
90 These
e results provve that the w
worse case, lowest natural frequenccy, occurs in Caveat #1: wh
hen the torq
que sensor aand the couppling springs are paired iinstead of th
he motor shaft a
m
and the coup
pling springss being paireed. It also sho
ows that thee only note worthy natur
w
ral frequencyy is mode tw
wo since it is closest to th
he forcing frrequency of 8 poles X 8000 RPM / 2 polles per rotattion = 320000 RPM. The llowest modee, is, in fact, well within the op
w
perating regime. The re
esulting amp
plitude plot o
of the mass at each mod
de is shown below (Figure 0‐17 and Figu
ure 0‐18). Th
hese plots sh
how that moode two has more vibration amplitud
de as th
he analysis m
moves down
n the length o
of the system
m. It must b
be noted how
wever that without dam
w
ping specifie
ed, the absolute magnituude of any m
mode is infin
nite, and thus caannot be compared. The
ere are two notes on thee following p
plots. First th
he scale of tthe x‐
axxis shows haalf masses w
which is incorrrect. This onnly appears due to the zzoom of the axis and can be iggnored for fu
uture mass p
plots. Secondd, the y‐axis displays Am
mplitude with
h a 1440 question marrk. This is to only to remind the readder that this value is for determiningg which caveat
w
s the best d
definition of the system to calculate natural freq
quency. Figure 0‐17: Caaveat #1 Mode
e Amplitudes ffor Each Masss in the Spring‐‐Mass System
Figure 0‐18: Caveat #2 Mod e Amplitudes of Each Mass 1441 The ke
ey take awayy from these
e plots is thaat moving th
he spring in tthe “caveat ##2” caase changed
d the shape o
of Mode 2 siignificantly. This does n
not make sen
nse physicallly; th
herefore Cavveat 2 is likely not representative off the actual ssystem. To invvestigate a different style sensor, thee Interface ttorque senso
or, modificattions needed to be
e made to the Matlab m‐‐file. The new
w mass‐spring diagram,, Figure 0‐19
9, hows the Intterface masss moment off inertias forr the drive and test side.. Interface uses a sh
te
echnology w
where a throu
ugh shaft haas a necked ddown portio
on in the mid
ddle and thiss measures the
m
e torque between the drrive and testt side. Accorrding to the Interface data sh
heet, the torrsional stiffn
ness is only o
of the neck ccomponent, thus the torrsional stiffn
ness of the shaft e
ends was calcculated assu
uming stainleess steel. Th
he moment o
of inertias arre different as w
well because
e the drive side has no m
measurement equipmentt attached whereas the t
w
test side of tthe sensor shaft has diffferent measurement equipment atttached. Figure
e 0‐19: Mass‐SSpring Diagram
m of Interface TTorque Sensorr with Caveat ##1 Shown (Cou
upler an
nd Torque Sensor Shaft Sprin
ngs Paired Toggether) Below
w is the set‐u
up for the Intterface m‐fille: %Mas
ss Moment
ts of Ine
ertia:
Jm = 6.271E-3; %kg*m
m^2 (moto
or) Artis
san GP150
0WC
Js = 1.295E-4; %kg*m
m^2 (moto
or shaft) assumin
ng stainl
less
stee
el
Jc = 4.5E-3; %kg*m^2
2 (couple
er) assum
ming R+W America BK2
coupler
c
1442 Jd = 1E-4; %kg*m^2 (driven/Unit Under Test (UUT) side
of torque sensor) assuming Interface T2 Torque sensor
Jt = 9E-5; %kg*m^2 (test/absorber side of torque
sensor)
%Torsional Stiffness (spring constants):
Ks = 1.33E5; %Nm/rad (motor shaft) calculated assuming
stainless steel
Kc = 1.91E5; %Nm/rad (coupler) given
Kt = 6.7E4; %Nm/rad (torque sensor) given
Kst = 1.977E5; %Nm/rad (torque sensor shaft) calculated
using stainless steel
% Relate the physical values to the generic spring mass
system
J1 = Jm+Js;
J2 = Jc;
J3 = Jd;
J4 = Jt;
J5 = Jc;
J6 = Jm+Js;
% Caveat #1: Assuming the coupler "spring" is on the
torque sensor shaft "spring"
% K1 = Ks;
% K2 = 1/(1/Kc + 1/Kst);
% K3 = Kt;
% K4 = 1/(1/Kc + 1/Kst);
% K5 = Ks;
143 % Ca
aveat #2: Assumin
ng the co
oupler "s
spring" i
is on the
e
motor
m
sha
aft "spri
ing"
K1 = 1/(1/Kc
c + 1/Ks);
K2 = Kst;
K3 = Kt;
K4 = Kst;
K5 = 1/(1/Kc
c + 1/Ks);
h caveat #1 aand caveat ##2, Table 0‐8
8 was producced After running the file for both
to
o critique the difference
e in assumptions. A plot of the modee amplitudess at each lo
ocation in the mass‐sprin
ng diagram ffrom Figure 0‐19 is show
wn in Figure 0‐20 and Figgure 0‐21. These figures are viisual represe
entations of the differen
nces at each point in thee syystem depen
nding on wh
hich caveat iss selected. Itt can again b
be determined that caveeat #1 is the worsst‐case scenario for earliest natural frequency.
Table 0‐8: Interface Sensor with R
R+W Couplers C
Caveat Compaarison Natural Frequencyy At: Mo
ode 2 Mo
ode 3 Mo
ode 4 Mo
ode 5 Mo
ode 6 Caveat #1 n (rpm
m) 20,140 67,3
320 72,8
880 308,,360 473,,060 Caveaat #2 n (rpm
m) 21,6640 51,7700 61,9900 439,3350 568,3330 1444 Differeence (#1‐##2) ‐1,50
00 15,62
20 10,98
80 ‐130,9
990 ‐95,270 Figure 0‐20: Detailed View
w of Caveat #1 w
with Only the First Three Mo
odes Shown Amplitude Plot t of Caveat #2 ffor Interface SSensor Figure 0‐21: Mode A
Lookin
ng at the different couplings availabble for the In
nterface set‐up and theirr natural frequency effectss on the systtem is anoth er valuable study. Show
wn below (Taable 1445 0‐9) is a table
e summarizin
ng these find
dings. The thhree differen
nt couplings with their re
espective J aand K terms were used for these calcculations. Figure 0‐22 an
nd Figure 0‐23 illustrate thesse findings for the first tthree modess. Table 0
0‐9: Interface ccoupling comp
parison resultss (all using caveeat#1) Natura
al Frequencyy At: Mo
ode 1 Mo
ode 2 Mo
ode 3 Mo
ode 4 Mo
ode 5 Mo
ode 6 R+W BK2 cou
R
upling n (rpm
m) 0 0
20,140 67,3
320 72,8
880 308,,360 473,,060 nterface coup
pling In
n (rpm) 0 23,4490 105,0090 125,1190 346,9990 495,3379 R++W BKM coup
pling n (rpm)
0
22,87
70 135,2
210 167,7
780 322,7
740 472,8
890 Figure 0‐22: Mode
e‐Amplitude Pllot of Caveat ##1 with the Intterface Sensor and the Supplied Interface Coup
pling 1446 Figu
ure 0‐23: Mode
e‐Amplitude Plot for Caveat #1 with Interfface Sensor an
nd BKM Couplings The next sensor to
o investigate
e is the natuural frequenccy of is the B
Binsfeld. Thiss se
ensor uses aa strain gaugge that is mo
ounted directtly on the m
motor shaft aand transmitts the data inductively. Figure 0
0‐24 pictorially illustrate s the differeent sizes of B
Binsfeld senssors all with a non
n‐contact collar. 1447 Figurre 0‐24: Binsfeeld Torque Sen
nsors
Since the Binsfeld
d attaches directly to thee shaft, only one coupler is needed aand th
he sensor’s inertia and sspring constaant can esseentially be ignored. Figurre 0‐25 show
ws how the masss‐spring mod
del is change
ed for analyssis. Figure
e 0‐25: Spring‐mass diagram of Binsfeld to
orque sensor seet‐up After running the Matlab code for this neew design, th
he followingg results seen
n in ere found. TThe results sh
how that thee second mo
ode has the highest yet seen Table 0‐10 we
natural frequency. 1448 Table 0‐10: Binsfeld natural frequency results Mode 2 Mode 3 Natural Frequency (rpm) 37,070 76,950 There are some assumptions to the Binsfeld model that need to be mentioned however. The model assumes that the mass of the strain gauge collar on the shaft is negligible because it is made of nylon and is mounted solidly on one shaft. The shaft that has the sensor collar mounted to it would have a higher mass moment of inertia but this was neglected for simplicity. A better approximation would be to add the moment of inertia of the sensor collar to the coupler inertia. The assumptions used to calculate these frequencies allow for the values to be higher than they would be in the real‐life application. Table 0‐11 summarizes the previous investigations into one table. Notice that the R+W BK2 and BKM couplers were each tried with different sensors for a more conclusive investigation. 149 Table 0‐11: Summarry of Natural FFrequency of D
Different Torqu
ue Sensors and
d Couplings (lo
owest re
eported naturaal frequency sh
hown for each
h case) 1550 Appendix E
A
E– Instructiions for Ch
hanging thee Absorberr Inverter ffrom Line SSync Control to Resolve
er Control ((GAS to Dyyno Set‐Up
p) 1. Unbollt cover (covver has dowe
el pins), discconnect fan o
on backside. 2. Cut th
he zip ties around the co
ontrol box (iff they were rreinstalled b
by the last usser) 3. Remo
ove the pin‐o
out harness ffrom the linee sync card
4. Unscrrew the big sscrews on eaach side of thhe yellow taabs and the sscrew holdin
ng the em
mpty alumin
num cover pllate on, to reelease the 88903‐CB main Card. Pull it out ve
ery carefully as the wirin
ng harness iss VERY shortt. Take out th
he line sync card by rem
moving the p
proper screw
ws. 1551 5. The litttle board is the feedbacck (resolver)) card adapteer. 6. Changge the wires from the lin
ne sync card to the feedb
back (resolvver) card adaapter accord
ding to this w
wiring spreaadsheet. Resolver Card
C
Pin
3
2
Internal Cable
ACK
BLA
WH
HT
Signal
V
0V
+15V
Wire Label
1
2
4
5
WH
HT
BR
RN
CH
HB
CH
H B-
11
1
1
12
6
WH
HT
CH
HA
7
1
RED
CH
H Ashield
14
1
1
15
1
16
1
17
1552 7. Insertt the pin‐outt into the ressolver card. Bend the wiires from thee screw termin
nals (there iss little clearaance for the ferrules to protrude without being bent) and install tthe DB14 fro
om the adap ter into the resolver carrd 8. Reinsttall resolver card into the 8903CB m
main card, maaking sure th
hat the conttacts line up properly. D
DO NOT ove
er tighten thee screws. Th
hey should be finger tigh
ht at the m
most. 1553 9. Slide ccontrol bay iinto its slotss. When you slide it in, taake your tim
me because it is very tight and the
e wires have little flex in them. 1554 10. Reinsttall the screw
ws from the aluminum ccover and th
he yellow tab
bs 11. Screw
w on the reso
olver card DB
B14 connecttion card 1555 12. Installl the fan, clo
ose the case with the pocket side covering the P
PC board allowiing adequate
e space for tthe fan, tigh ten the casee bolts not very tight. ` 1556